Merge remote-tracking branch 'origin/main' into alexcheema/prefill-progress

chore: fix formatting after merge
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-18 14:55:13 -05:00 · 2026-02-13 09:37:55 -08:00 · 2026-02-13 06:54:18 -08:00 · 2026-02-13 05:58:45 -08:00 · 2026-02-05 05:38:14 -08:00 · 2026-02-05 05:30:22 -08:00
74 changed files with 3072 additions and 2253 deletions
--- a/.github/workflows/pipeline.yml
+++ b/.github/workflows/pipeline.yml
@@ -8,6 +8,33 @@ on:
      - main

 jobs:
+  typecheck:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          lfs: false
+
+      - uses: cachix/install-nix-action@v31
+        with:
+          nix_path: nixpkgs=channel:nixos-unstable
+
+      - uses: cachix/cachix-action@v14
+        name: Configure Cachix
+        with:
+          name: exo
+          authToken: "${{ secrets.CACHIX_AUTH_TOKEN }}"
+
+      - name: Load nix develop environment
+        run: nix run github:nicknovitski/nix-develop/v1
+
+      - name: Sync dependencies
+        run: uv sync --all-packages
+
+      - name: Run type checker
+        run: uv run basedpyright --project pyproject.toml
+
  nix:
    name: Build and check (${{ matrix.system }})
    runs-on: ${{ matrix.runner }}
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -125,9 +125,9 @@ dependencies = [

 [[package]]
 name = "anyhow"
-version = "1.0.101"
+version = "1.0.100"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5f0e0fee31ef5ed1ba1316088939cea399010ed7731dba877ed44aeb407a75ea"
+checksum = "a23eb6b1614318a8071c9b2521f36b424b2c83db5eb3a0fead4a6c0809af6e61"

 [[package]]
 name = "arc-swap"
@@ -141,6 +141,12 @@ version = "0.3.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "76a2e8124351fda1ef8aaaa3bbd7ebbcb486bbcd4225aca0aa0d84bb2db8fecb"

+[[package]]
+name = "arrayvec"
+version = "0.7.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7c02d123df017efcdfbd739ef81735b36c5ba83ec3c59c80a9d7ecc718f92e50"
+
 [[package]]
 name = "asn1-rs"
 version = "0.7.1"
@@ -165,7 +171,7 @@ checksum = "3109e49b1e4909e9db6515a30c633684d68cdeaa252f215214cb4fa1a5bfee2c"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 "synstructure",
 ]

@@ -177,7 +183,7 @@ checksum = "7b18050c2cd6fe86c3a76584ef5e0baf286d038cda203eb6223df2cc413565f7"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -224,7 +230,7 @@ checksum = "9035ad2d096bed7955a320ee7e2230574d28fd3c3a0f186cbea1ff3c7eed5dbb"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -298,6 +304,19 @@ version = "1.8.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "55248b47b0caf0546f7988906588779981c43bb1bc9d0c44087278f80cdb44ba"

+[[package]]
+name = "bigdecimal"
+version = "0.4.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "560f42649de9fa436b73517378a147ec21f6c997a546581df4b4b31677828934"
+dependencies = [
+ "autocfg",
+ "libm",
+ "num-bigint",
+ "num-integer",
+ "num-traits",
+]
+
 [[package]]
 name = "bimap"
 version = "0.6.3"
@@ -421,9 +440,9 @@ dependencies = [

 [[package]]
 name = "chrono"
-version = "0.4.43"
+version = "0.4.42"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fac4744fb15ae8337dc853fee7fb3f4e48c0fbaa23d0afe49c447b4fab126118"
+checksum = "145052bdd345b87320e369255277e3fb5152762ad123a901ef5c262dd38fe8d2"
 dependencies = [
 "iana-time-zone",
 "js-sys",
@@ -497,6 +516,15 @@ version = "0.4.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "2f421161cb492475f1661ddc9815a745a1c894592070661180fdec3d4872e9c3"

+[[package]]
+name = "convert_case"
+version = "0.10.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "633458d4ef8c78b72454de2d54fd6ab2e60f9e02be22f3c6104cdc8a4e0fceb9"
+dependencies = [
+ "unicode-segmentation",
+]
+
 [[package]]
 name = "core-foundation"
 version = "0.9.4"
@@ -644,7 +672,7 @@ checksum = "f46882e17999c6cc590af592290432be3bce0428cb0d5f8b6715e4dc7b383eb3"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -670,7 +698,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "8d162beedaa69905488a8da94f5ac3edb4dd4788b732fadb7bd120b2625c1976"
 dependencies = [
 "data-encoding",
- "syn 1.0.109",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -681,7 +709,7 @@ checksum = "780eb241654bf097afb00fc5f054a09b687dad862e485fdcf8399bb056565370"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -718,6 +746,29 @@ dependencies = [
 "powerfmt",
 ]

+[[package]]
+name = "derive_more"
+version = "2.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "10b768e943bed7bf2cab53df09f4bc34bfd217cdb57d971e769874c9a6710618"
+dependencies = [
+ "derive_more-impl",
+]
+
+[[package]]
+name = "derive_more-impl"
+version = "2.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6d286bfdaf75e988b4a78e013ecd79c581e06399ab53fbacd2d916c2f904f30b"
+dependencies = [
+ "convert_case",
+ "proc-macro2",
+ "quote",
+ "rustc_version",
+ "syn 2.0.111",
+ "unicode-xid",
+]
+
 [[package]]
 name = "digest"
 version = "0.10.7"
@@ -738,7 +789,7 @@ checksum = "97369cbbc041bc366949bc74d34658d6cda5621039731c6310521892a3a20ae0"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -820,7 +871,7 @@ dependencies = [
 "heck",
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -888,17 +939,22 @@ name = "exo_pyo3_bindings"
 version = "0.0.1"
 dependencies = [
 "delegate",
+ "derive_more",
 "env_logger",
 "extend",
- "futures-lite",
+ "futures",
+ "impl-trait-for-tuples",
 "libp2p",
 "log",
 "networking",
+ "once_cell",
 "pin-project",
 "pyo3",
 "pyo3-async-runtimes",
 "pyo3-log",
 "pyo3-stub-gen",
+ "thiserror 2.0.17",
+ "thread_local",
 "tokio",
 "util",
 ]
@@ -911,15 +967,9 @@ checksum = "311a6d2f1f9d60bff73d2c78a0af97ed27f79672f15c238192a5bbb64db56d00"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

-[[package]]
-name = "fastrand"
-version = "2.3.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "37909eebbb50d72f9059c3b6d82c0463f2ff062c9e95845c43a6c9c0355411be"
-
 [[package]]
 name = "ff"
 version = "0.13.1"
@@ -1028,10 +1078,7 @@ version = "2.6.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "f78e10609fe0e0b3f4157ffab1876319b5b0db102a2c60dc4626306dc46b44ad"
 dependencies = [
- "fastrand",
 "futures-core",
- "futures-io",
- "parking",
 "pin-project-lite",
 ]

@@ -1043,7 +1090,7 @@ checksum = "162ee34ebcb7c64a8abebc059ce0fee27c2262618d7b60ed8faf72fef13c3650"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -1593,6 +1640,17 @@ dependencies = [
 "xmltree",
 ]

+[[package]]
+name = "impl-trait-for-tuples"
+version = "0.2.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a0eb5a3343abf848c0984fe4604b2b105da9539376e24fc0a3b0007411ae4fd9"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn 2.0.111",
+]
+
 [[package]]
 name = "indexmap"
 version = "2.12.1"
@@ -1657,7 +1715,7 @@ dependencies = [
 "heck",
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -1711,7 +1769,7 @@ checksum = "980af8b43c3ad5d8d349ace167ec8170839f753a42d233ba19e08afe1850fa69"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -1771,6 +1829,12 @@ version = "0.2.178"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "37c93d8daa9d8a012fd8ab92f088405fb202ea0b6ab73ee2482ae66af4f42091"

+[[package]]
+name = "libm"
+version = "0.2.15"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f9fbbcab51052fe104eb5e5d351cf728d30a5be1fe14d9be8a3b097481fb97de"
+
 [[package]]
 name = "libp2p"
 version = "0.56.0"
@@ -2300,7 +2364,7 @@ checksum = "dd297cf53f0cb3dee4d2620bb319ae47ef27c702684309f682bdb7e55a18ae9c"
 dependencies = [
 "heck",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -2760,13 +2824,16 @@ name = "networking"
 version = "0.0.1"
 dependencies = [
 "delegate",
+ "derive_more",
 "either",
 "extend",
- "futures-lite",
+ "futures",
 "futures-timer",
+ "impl-trait-for-tuples",
 "keccak-const",
 "libp2p",
 "log",
+ "thiserror 2.0.17",
 "tokio",
 "tracing-subscriber",
 "util",
@@ -2838,9 +2905,9 @@ dependencies = [

 [[package]]
 name = "num-conv"
-version = "0.2.0"
+version = "0.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cf97ec579c3c42f953ef76dbf8d55ac91fb219dde70e49aa4a6b7d74e9919050"
+checksum = "51d515d32fb182ee37cda2ccdcb92950d6a3c2893aa280e540671c2cd0f3b1d9"

 [[package]]
 name = "num-integer"
@@ -2851,6 +2918,17 @@ dependencies = [
 "num-traits",
 ]

+[[package]]
+name = "num-rational"
+version = "0.4.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f83d14da390562dca69fc84082e73e548e1ad308d24accdedd2720017cb37824"
+dependencies = [
+ "num-bigint",
+ "num-integer",
+ "num-traits",
+]
+
 [[package]]
 name = "num-traits"
 version = "0.2.19"
@@ -3053,7 +3131,7 @@ checksum = "6e918e4ff8c4549eb882f14b3a4bc8c8bc93de829416eacf579f1207a8fbf861"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -3165,9 +3243,9 @@ dependencies = [

 [[package]]
 name = "proc-macro2"
-version = "1.0.106"
+version = "1.0.103"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8fd00f0bb2e90d81d1044c2b32617f68fcb9fa3bb7640c23e9c748e53fb30934"
+checksum = "5ee95bc4ef87b8d5ba32e8b7714ccc834865276eab0aed5c9958d00ec45f49e8"
 dependencies = [
 "unicode-ident",
 ]
@@ -3192,7 +3270,7 @@ checksum = "440f724eba9f6996b75d63681b0a92b06947f1457076d503a4d2e2c8f56442b8"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -3201,14 +3279,28 @@ version = "0.27.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "ab53c047fcd1a1d2a8820fe84f05d6be69e9526be40cb03b73f86b6b03e6d87d"
 dependencies = [
+ "bigdecimal",
+ "either",
+ "hashbrown 0.16.1",
+ "indexmap",
 "indoc",
+ "inventory",
 "libc",
+ "lock_api",
 "memoffset",
+ "num-bigint",
+ "num-complex",
+ "num-rational",
+ "num-traits",
 "once_cell",
+ "ordered-float",
+ "parking_lot",
 "portable-atomic",
 "pyo3-build-config",
 "pyo3-ffi",
 "pyo3-macros",
+ "rust_decimal",
+ "smallvec",
 "unindent",
 ]

@@ -3236,7 +3328,7 @@ checksum = "bcd7d70ee0ca1661c40407e6f84e4463ef2658c90a9e2fbbd4515b2bcdfcaeca"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -3278,7 +3370,7 @@ dependencies = [
 "proc-macro2",
 "pyo3-macros-backend",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -3291,14 +3383,14 @@ dependencies = [
 "proc-macro2",
 "pyo3-build-config",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
 name = "pyo3-stub-gen"
-version = "0.19.0"
+version = "0.17.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b159f7704044f57d058f528a6f1f22a0a0a327dcb595c5fb38beae658e0338d6"
+checksum = "398b833826a83ca72c1e26d1b2c7c71f9ca7c3bfc74eacc663901895c362ae33"
 dependencies = [
 "anyhow",
 "chrono",
@@ -3313,25 +3405,22 @@ dependencies = [
 "ordered-float",
 "pyo3",
 "pyo3-stub-gen-derive",
- "rustpython-parser",
 "serde",
- "serde_json",
- "time",
 "toml",
 ]

 [[package]]
 name = "pyo3-stub-gen-derive"
-version = "0.19.0"
+version = "0.17.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a8c79e7c5b1fcec7c39ab186594658a971c59911eb6fbab5a5932cf2318534be"
+checksum = "2426ba759d848787239d80f9fdb1f223786976f87fb6c3da8188ca7c17744b28"
 dependencies = [
 "heck",
 "indexmap",
 "proc-macro2",
 "quote",
 "rustpython-parser",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -3414,9 +3503,9 @@ dependencies = [

 [[package]]
 name = "quote"
-version = "1.0.44"
+version = "1.0.42"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "21b2ebcf727b7760c461f091f9f0f539b77b8e87f2fd88131e7f1b433b3cece4"
+checksum = "a338cc41d27e6cc6dce6cefc13a0729dfbb81c262b1f519331575dd80ef3067f"
 dependencies = [
 "proc-macro2",
 ]
@@ -3652,6 +3741,16 @@ dependencies = [
 "tokio",
 ]

+[[package]]
+name = "rust_decimal"
+version = "1.39.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "35affe401787a9bd846712274d97654355d21b2a2c092a3139aabe31e9022282"
+dependencies = [
+ "arrayvec",
+ "num-traits",
+]
+
 [[package]]
 name = "rustc-hash"
 version = "1.1.0"
@@ -3887,7 +3986,7 @@ checksum = "d540f220d3187173da220f885ab66608367b6574e925011a9353e4badda91d79"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -3905,9 +4004,9 @@ dependencies = [

 [[package]]
 name = "serde_spanned"
-version = "1.0.4"
+version = "1.0.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f8bbf91e5a4d6315eee45e704372590b30e260ee83af6639d64557f51b067776"
+checksum = "e24345aa0fe688594e73770a5f6d1b216508b4f93484c0026d521acd30134392"
 dependencies = [
 "serde_core",
 ]
@@ -4095,9 +4194,9 @@ dependencies = [

 [[package]]
 name = "syn"
-version = "2.0.116"
+version = "2.0.111"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3df424c70518695237746f84cede799c9c58fcb37450d7b23716568cc8bc69cb"
+checksum = "390cc9a294ab71bdb1aa2e99d13be9c753cd2d7bd6560c77118597410c4d2e87"
 dependencies = [
 "proc-macro2",
 "quote",
@@ -4112,7 +4211,7 @@ checksum = "728a70f3dbaf5bab7f0c4b1ac8d7ae5ea60a4b5549c8a5914361c99147a709d2"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -4188,7 +4287,7 @@ checksum = "4fee6c4efc90059e10f81e6d42c60a18f76588c3d74cb83a0b242a2b6c7504c1"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -4199,7 +4298,7 @@ checksum = "3ff15c8ecd7de3849db632e14d18d2571fa09dfc5ed93479bc4485c7a517c913"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -4213,30 +4312,30 @@ dependencies = [

 [[package]]
 name = "time"
-version = "0.3.47"
+version = "0.3.44"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "743bd48c283afc0388f9b8827b976905fb217ad9e647fae3a379a9283c4def2c"
+checksum = "91e7d9e3bb61134e77bde20dd4825b97c010155709965fedf0f49bb138e52a9d"
 dependencies = [
 "deranged",
 "itoa",
 "num-conv",
 "powerfmt",
- "serde_core",
+ "serde",
 "time-core",
 "time-macros",
 ]

 [[package]]
 name = "time-core"
-version = "0.1.8"
+version = "0.1.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7694e1cfe791f8d31026952abf09c69ca6f6fa4e1a1229e18988f06a04a12dca"
+checksum = "40868e7c1d2f0b8d73e4a8c7f0ff63af4f6d19be117e90bd73eb1d62cf831c6b"

 [[package]]
 name = "time-macros"
-version = "0.2.27"
+version = "0.2.24"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2e70e4c5a0e0a8a4823ad65dfe1a6930e4f4d756dcd9dd7939022b5e8c501215"
+checksum = "30cfb0125f12d9c277f35663a0a33f8c30190f4e4574868a330595412d34ebf3"
 dependencies = [
 "num-conv",
 "time-core",
@@ -4312,7 +4411,7 @@ checksum = "af407857209536a95c8e56f8231ef2c2e2aff839b22e07a1ffcbc617e9db9fa5"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -4330,9 +4429,9 @@ dependencies = [

 [[package]]
 name = "toml"
-version = "1.0.2+spec-1.1.0"
+version = "0.9.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d1dfefef6a142e93f346b64c160934eb13b5594b84ab378133ac6815cb2bd57f"
+checksum = "f0dc8b1fb61449e27716ec0e1bdf0f6b8f3e8f6b05391e8497b8b6d7804ea6d8"
 dependencies = [
 "indexmap",
 "serde_core",
@@ -4345,27 +4444,27 @@ dependencies = [

 [[package]]
 name = "toml_datetime"
-version = "1.0.0+spec-1.1.0"
+version = "0.7.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "32c2555c699578a4f59f0cc68e5116c8d7cabbd45e1409b989d4be085b53f13e"
+checksum = "f2cdb639ebbc97961c51720f858597f7f24c4fc295327923af55b74c3c724533"
 dependencies = [
 "serde_core",
 ]

 [[package]]
 name = "toml_parser"
-version = "1.0.9+spec-1.1.0"
+version = "1.0.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "702d4415e08923e7e1ef96cd5727c0dfed80b4d2fa25db9647fe5eb6f7c5a4c4"
+checksum = "c0cbe268d35bdb4bb5a56a2de88d0ad0eb70af5384a99d648cd4b3d04039800e"
 dependencies = [
 "winnow",
 ]

 [[package]]
 name = "toml_writer"
-version = "1.0.6+spec-1.1.0"
+version = "1.0.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ab16f14aed21ee8bfd8ec22513f7287cd4a91aa92e44edfe2c17ddd004e92607"
+checksum = "df8b2b54733674ad286d16267dcfc7a71ed5c776e4ac7aa3c3e2561f7c637bf2"

 [[package]]
 name = "tower-service"
@@ -4392,7 +4491,7 @@ checksum = "7490cfa5ec963746568740651ac6781f701c9c5ea257c58e057f3ba8cf69e8da"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -4516,12 +4615,24 @@ version = "1.0.22"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "9312f7c4f6ff9069b165498234ce8be658059c6728633667c526e27dc2cf1df5"

+[[package]]
+name = "unicode-segmentation"
+version = "1.12.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f6ccf251212114b54433ec949fd6a7841275f9ada20dddd2f29e9ceea4501493"
+
 [[package]]
 name = "unicode-width"
 version = "0.2.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b4ac048d71ede7ee76d585517add45da530660ef4390e49b098733c6e897f254"

+[[package]]
+name = "unicode-xid"
+version = "0.2.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ebc1c04c71510c7f702b52b7c350734c9ff1295c464a03335b00bb84fc54f853"
+
 [[package]]
 name = "unicode_names2"
 version = "1.3.0"
@@ -4704,7 +4815,7 @@ dependencies = [
 "bumpalo",
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 "wasm-bindgen-shared",
 ]

@@ -4846,7 +4957,7 @@ checksum = "9107ddc059d5b6fbfbffdfa7a7fe3e22a226def0b2608f72e9d552763d3e1ad7"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -4857,7 +4968,7 @@ checksum = "053e2e040ab57b9dc951b72c264860db7eb3b0200ba345b4e4c3b14f67855ddf"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -4868,7 +4979,7 @@ checksum = "29bee4b38ea3cde66011baa44dba677c432a78593e202392d1e9070cf2a7fca7"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -4879,7 +4990,7 @@ checksum = "3f316c4a2570ba26bbec722032c4099d8c8bc095efccdc15688708623367e358"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -5268,7 +5379,7 @@ checksum = "b659052874eb698efe5b9e8cf382204678a0086ebf46982b79d6ca3182927e5d"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 "synstructure",
 ]

@@ -5289,7 +5400,7 @@ checksum = "d8a8d209fdf45cf5138cbb5a506f6b52522a25afccc534d1475dad8e31105c6a"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -5309,7 +5420,7 @@ checksum = "d71e5d6e06ab090c67b5e44993ec16b72dcbaabc526db883a360057678b48502"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 "synstructure",
 ]

@@ -5330,7 +5441,7 @@ checksum = "ce36e65b0d2999d2aafac989fb249189a141aee1f53c612c1f37d72631959f69"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]

 [[package]]
@@ -5363,5 +5474,5 @@ checksum = "eadce39539ca5cb3985590102671f2567e659fca9666581ad3411d59207951f3"
 dependencies = [
 "proc-macro2",
 "quote",
- "syn 2.0.116",
+ "syn 2.0.111",
 ]
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -26,22 +26,49 @@ opt-level = 3
 networking = { path = "rust/networking" }
 util = { path = "rust/util" }

+# Proc-macro authoring tools
+syn = "2.0"
+quote = "1.0"
+proc-macro2 = "1.0"
+darling = "0.20"
+
 # Macro dependecies
 extend = "1.2"
 delegate = "0.13"
+impl-trait-for-tuples = "0.2"
+clap = "4.5"
+derive_more = { version = "2.0.1", features = ["display"] }
 pin-project = "1"

 # Utility dependencies
+itertools = "0.14"
+thiserror = "2"
+internment = "0.8"
+recursion = "0.5"
+regex = "1.11"
+once_cell = "1.21"
+thread_local = "1.1"
+bon = "3.4"
+generativity = "1.1"
+anyhow = "1.0"
 keccak-const = "0.2"

+# Functional generics/lenses frameworks
+frunk_core = "0.4"
+frunk = "0.4"
+frunk_utils = "0.2"
+frunk-enum-core = "0.3"
+
 # Async dependencies
 tokio = "1.46"
 futures = "0.3"
-futures-lite = "2.6.1"
+futures-util = "0.3"
 futures-timer = "3.0"

 # Data structures
 either = "1.15"
+ordered-float = "5.0"
+ahash = "0.8"

 # Tracing/logging
 log = "0.4"
--- a/MISSED_THINGS.md
+++ b/MISSED_THINGS.md
@@ -5,21 +5,21 @@
 [X] Fetching download status of all models on start
 [X] Deduplication of tasks in plan_step.
 [X] resolve_allow_patterns should just be wildcard now.
-[X] no mx_barrier in genreate.py mlx_generate at the end.
+[] no mx_barrier in genreate.py mlx_generate at the end.
 [] cache assertion not needed in auto_parallel.py PipelineLastLayer.
-[X] GPTOSS support dropped in auto_parallel.py.
-[X] sharding changed "all-to-sharded" became _all_to_sharded in auto_parallel.py.
-[X] same as above with "sharded-to-all" became _sharded_to_all in auto_parallel.py.
-[X] Dropped support for Ministral3Model, DeepseekV32Model, Glm4MoeModel, Qwen3NextModel, GptOssMode in auto_parallel.py.
+[] GPTOSS support dropped in auto_parallel.py.
+[] sharding changed "all-to-sharded" became _all_to_sharded in auto_parallel.py.
+[] same as above with "sharded-to-all" became _sharded_to_all in auto_parallel.py.
+[] Dropped support for Ministral3Model, DeepseekV32Model, Glm4MoeModel, Qwen3NextModel, GptOssMode in auto_parallel.py.
 [] Dropped prefill/decode code in auto_parallel.py and utils_mlx.py.
 [X] KV_CACHE_BITS should be None to disable quantized KV cache.
-[X] Dropped _set_nofile_limit in utils_mlx.py.
-[X] We have group optional in load_mlx_items in utils_mlx.py.
-[X] Dropped add_missing_chat_templates for GptOss in load_mlx_items in utils_mlx.py.
-[X] Dropped model.make_cache in make_kv_cache in utils_mlx.py.
+[] Dropped _set_nofile_limit in utils_mlx.py.
+[] We have group optional in load_mlx_items in utils_mlx.py.
+[] Dropped add_missing_chat_templates for GptOss in load_mlx_items in utils_mlx.py.
+[] Dropped model.make_cache in make_kv_cache in utils_mlx.py.
 [X] We put cache limit back in utils_mlx.py.
-[X] topology.py remove_node removes the connections after checking if node is is in self._node_id_to_rx_id_map. on beta_1 it checks after, so would remove stale connections I guess?
-[X] Missing Glm 4.7 model cards (this isn't ready yet but should be picked up, probably create an issue... the blocker is transforemrs version doesn't support the tokenizer for Glm 4.7. rc-1 does but we can't upgrade as it breaks other things.)
+[] topology.py remove_node removes the connections after checking if node is is in self._node_id_to_rx_id_map. on beta_1 it checks after, so would remove stale connections I guess?
+[] Missing Glm 4.7 model cards (this isn't ready yet but should be picked up, probably create an issue... the blocker is transforemrs version doesn't support the tokenizer for Glm 4.7. rc-1 does but we can't upgrade as it breaks other things.)
 [] try-except in _command_processor only excepts ValueError. This was silently failing leading to un-debuggable errors (we had a KeyError that was happening ). Changed this to catch Exception instead of ValueError. See exo-v2 89ae38405e0052e3c22405daf094b065878aa873 and fb99fea69b5a39017efc90c5dad0072e677455f0.
 [X] In placement.py, place_instance no longer looks at model_meta.supports_tensor and check if this tensor parallel number of nodes is supported by the model's tensor dimensions.
 [X] In placement.py, place_instanec, we no longer have the special case to exclude DeepSeek v3.1 pipeline parallel (it doesn't work).
--- a/README.md
+++ b/README.md
@@ -72,23 +72,16 @@ There are two ways to run exo:

 ### Run from Source (macOS)

-If you have [Nix](https://nixos.org/) installed, you can skip most of the steps below and run exo directly (after accepting the Cachix cache):
-
-```bash
-nix run .#exo
-```
-
 **Prerequisites:**
- [Xcode](https://developer.apple.com/xcode/) (provides the Metal ToolChain required for MLX compilation)
 - [brew](https://github.com/Homebrew/brew) (for simple package management on macOS)
-
+  
  ```bash
  /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
  ```
 - [uv](https://github.com/astral-sh/uv) (for Python dependency management)
 - [macmon](https://github.com/vladkens/macmon) (for hardware monitoring on Apple Silicon)
 - [node](https://github.com/nodejs/node) (for building the dashboard)
-
+  
  ```bash
  brew install uv macmon node
  ```
--- a/app/EXO/EXO/ExoProcessController.swift
+++ b/app/EXO/EXO/ExoProcessController.swift
@@ -126,37 +126,11 @@ final class ExoProcessController: ObservableObject {
            return
        }
        process.terminationHandler = nil
-        status = .stopped
-
-        guard process.isRunning else {
-            self.process = nil
-            return
+        if process.isRunning {
+            process.terminate()
        }
-
-        let proc = process
        self.process = nil
-
-        Task.detached {
-            proc.interrupt()
-
-            for _ in 0..<50 {
-                if !proc.isRunning { return }
-                try? await Task.sleep(nanoseconds: 100_000_000)
-            }
-
-            if proc.isRunning {
-                proc.terminate()
-            }
-
-            for _ in 0..<30 {
-                if !proc.isRunning { return }
-                try? await Task.sleep(nanoseconds: 100_000_000)
-            }
-
-            if proc.isRunning {
-                kill(proc.processIdentifier, SIGKILL)
-            }
-        }
+        status = .stopped
    }

    func restart() {
--- a/bench/bench.toml
+++ b/bench/bench.toml
@@ -1,7 +0,0 @@
-# Canary benchmark manifest
-#
-# Lists the suite files to include. Each file defines benchmarks
-# with shared constraints, topology, and default args.
-include = [
-    "single-m3-ultra.toml",
-]
--- a/bench/exo_bench.py
+++ b/bench/exo_bench.py
@@ -288,151 +288,6 @@ def resolve_model_short_id(client: ExoClient, model_arg: str) -> tuple[str, str]
    raise ValueError(f"Model not found in /models: {model_arg}")


-def run_planning_phase(
-    client: ExoClient,
-    full_model_id: str,
-    preview: dict[str, Any],
-    danger_delete: bool,
-    timeout: float,
-    settle_deadline: float | None,
-) -> None:
-    """Check disk space and ensure model is downloaded before benchmarking."""
-    # Get model size from /models
-    models = client.request_json("GET", "/models") or {}
-    model_bytes = 0
-    for m in models.get("data", []):
-        if m.get("hugging_face_id") == full_model_id:
-            model_bytes = m.get("storage_size_megabytes", 0) * 1024 * 1024
-            break
-
-    if not model_bytes:
-        logger.warning(
-            f"Could not determine size for {full_model_id}, skipping disk check"
-        )
-        return
-
-    # Get nodes from preview
-    inner = unwrap_instance(preview["instance"])
-    node_ids = list(inner["shardAssignments"]["nodeToRunner"].keys())
-    runner_to_shard = inner["shardAssignments"]["runnerToShard"]
-
-    state = client.request_json("GET", "/state")
-    downloads = state.get("downloads", {})
-    node_disk = state.get("nodeDisk", {})
-
-    for node_id in node_ids:
-        node_downloads = downloads.get(node_id, [])
-
-        # Check if model already downloaded on this node
-        already_downloaded = any(
-            "DownloadCompleted" in p
-            and unwrap_instance(p["DownloadCompleted"]["shardMetadata"])["modelCard"][
-                "modelId"
-            ]
-            == full_model_id
-            for p in node_downloads
-        )
-        if already_downloaded:
-            continue
-
-        # Wait for disk info if settle_deadline is set
-        disk_info = node_disk.get(node_id, {})
-        backoff = _SETTLE_INITIAL_BACKOFF_S
-        while not disk_info and settle_deadline and time.monotonic() < settle_deadline:
-            remaining = settle_deadline - time.monotonic()
-            logger.info(
-                f"Waiting for disk info on {node_id} ({remaining:.0f}s remaining)..."
-            )
-            time.sleep(min(backoff, remaining))
-            backoff = min(backoff * _SETTLE_BACKOFF_MULTIPLIER, _SETTLE_MAX_BACKOFF_S)
-            state = client.request_json("GET", "/state")
-            node_disk = state.get("nodeDisk", {})
-            disk_info = node_disk.get(node_id, {})
-
-        if not disk_info:
-            logger.warning(f"No disk info for {node_id}, skipping space check")
-            continue
-
-        avail = disk_info.get("available", {}).get("inBytes", 0)
-        if avail >= model_bytes:
-            continue
-
-        if not danger_delete:
-            raise RuntimeError(
-                f"Insufficient disk on {node_id}: need {model_bytes // (1024**3)}GB, "
-                f"have {avail // (1024**3)}GB. Use --danger-delete-downloads to free space."
-            )
-
-        # Delete from smallest to largest
-        completed = [
-            (
-                unwrap_instance(p["DownloadCompleted"]["shardMetadata"])["modelCard"][
-                    "modelId"
-                ],
-                p["DownloadCompleted"]["totalBytes"]["inBytes"],
-            )
-            for p in node_downloads
-            if "DownloadCompleted" in p
-        ]
-        for del_model, size in sorted(completed, key=lambda x: x[1]):
-            logger.info(f"Deleting {del_model} from {node_id} ({size // (1024**2)}MB)")
-            client.request_json("DELETE", f"/download/{node_id}/{del_model}")
-            avail += size
-            if avail >= model_bytes:
-                break
-
-        if avail < model_bytes:
-            raise RuntimeError(f"Could not free enough space on {node_id}")
-
-    # Start downloads (idempotent)
-    for node_id in node_ids:
-        runner_id = inner["shardAssignments"]["nodeToRunner"][node_id]
-        shard = runner_to_shard[runner_id]
-        client.request_json(
-            "POST",
-            "/download/start",
-            body={
-                "targetNodeId": node_id,
-                "shardMetadata": shard,
-            },
-        )
-        logger.info(f"Started download on {node_id}")
-
-    # Wait for downloads
-    start = time.time()
-    while time.time() - start < timeout:
-        state = client.request_json("GET", "/state")
-        downloads = state.get("downloads", {})
-        all_done = True
-        for node_id in node_ids:
-            done = any(
-                "DownloadCompleted" in p
-                and unwrap_instance(p["DownloadCompleted"]["shardMetadata"])[
-                    "modelCard"
-                ]["modelId"]
-                == full_model_id
-                for p in downloads.get(node_id, [])
-            )
-            failed = [
-                p["DownloadFailed"]["errorMessage"]
-                for p in downloads.get(node_id, [])
-                if "DownloadFailed" in p
-                and unwrap_instance(p["DownloadFailed"]["shardMetadata"])["modelCard"][
-                    "modelId"
-                ]
-                == full_model_id
-            ]
-            if failed:
-                raise RuntimeError(f"Download failed on {node_id}: {failed[0]}")
-            if not done:
-                all_done = False
-        if all_done:
-            return
-        time.sleep(1)
-
-    raise TimeoutError("Downloads did not complete in time")
-
-
 def placement_filter(instance_meta: str, wanted: str) -> bool:
    s = (instance_meta or "").lower()
    if wanted == "both":
@@ -680,11 +535,6 @@ def main() -> int:
        default=0,
        help="Max seconds to wait for the cluster to produce valid placements (0 = try once).",
    )
-    ap.add_argument(
-        "--danger-delete-downloads",
-        action="store_true",
-        help="Delete existing models from smallest to largest to make room for benchmark model.",
-    )
    args = ap.parse_args()

    pp_list = parse_int_list(args.pp)
@@ -719,16 +569,13 @@ def main() -> int:
        logger.error("[exo-bench] tokenizer usable but prompt sizing failed")
        raise

-    settle_deadline = (
-        time.monotonic() + args.settle_timeout if args.settle_timeout > 0 else None
-    )
-
    selected = fetch_and_filter_placements(client, full_model_id, args)

-    if not selected and settle_deadline:
+    if not selected and args.settle_timeout > 0:
        backoff = _SETTLE_INITIAL_BACKOFF_S
-        while not selected and time.monotonic() < settle_deadline:
-            remaining = settle_deadline - time.monotonic()
+        deadline = time.monotonic() + args.settle_timeout
+        while not selected and time.monotonic() < deadline:
+            remaining = deadline - time.monotonic()
            logger.warning(
                f"No valid placements yet (cluster may still be settling). "
                f"Retrying in {backoff:.1f}s ({remaining:.0f}s remaining)..."
@@ -760,16 +607,6 @@ def main() -> int:
    if args.dry_run:
        return 0

-    logger.info("Planning phase: checking downloads...")
-    run_planning_phase(
-        client,
-        full_model_id,
-        selected[0],
-        args.danger_delete_downloads,
-        args.timeout,
-        settle_deadline,
-    )
-
    all_rows: list[dict[str, Any]] = []

    for preview in selected:
--- a/bench/single-m3-ultra.toml
+++ b/bench/single-m3-ultra.toml
@@ -1,189 +0,0 @@
-# Single-node M3 Ultra benchmarks
-#
-# Shared constraints applied to ALL benchmarks in this file.
-constraints = [
-    "All(MacOsBuild(=25D125))",
-    "Hosts(=1)",
-    "All(Chip(m3_ultra))",
-    "All(GpuCores(=80))",
-]
-
-[topology]
-type = "none"
-
-# Default args merged into each benchmark's args (benchmark-level args win).
-[defaults]
-pp = [512, 2048, 8192, 16384]
-tg = 128
-
-[[benchmark]]
-model = "mlx-community/Meta-Llama-3.1-70B-Instruct-4bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/gpt-oss-120b-MXFP4-Q8"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/GLM-4.7-Flash-8bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-Coder-Next-6bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-30B-A3B-8bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-0.6B-4bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-0.6B-8bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Llama-3.2-1B-Instruct-4bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Llama-3.2-3B-Instruct-4bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Llama-3.2-3B-Instruct-8bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Meta-Llama-3.1-8B-Instruct-4bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Meta-Llama-3.1-8B-Instruct-8bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Meta-Llama-3.1-8B-Instruct-bf16"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/gpt-oss-20b-MXFP4-Q8"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-30B-A3B-4bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/GLM-4.7-Flash-4bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/GLM-4.7-Flash-5bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/GLM-4.7-Flash-6bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Llama-3.3-70B-Instruct-4bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-Coder-Next-4bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-Coder-Next-5bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-Coder-Next-8bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-Next-80B-A3B-Instruct-4bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-Next-80B-A3B-Instruct-8bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-Next-80B-A3B-Thinking-4bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-Next-80B-A3B-Thinking-8bit"
-extra_constraints = ["All(Memory(>=96GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Llama-3.3-70B-Instruct-8bit"
-extra_constraints = ["All(Memory(>=256GiB))"]
-
-[[benchmark]]
-model = "mlx-community/llama-3.3-70b-instruct-fp16"
-extra_constraints = ["All(Memory(>=256GiB))"]
-
-[[benchmark]]
-model = "mlx-community/GLM-4.5-Air-8bit"
-extra_constraints = ["All(Memory(>=256GiB))"]
-
-[[benchmark]]
-model = "mlx-community/GLM-4.5-Air-bf16"
-extra_constraints = ["All(Memory(>=256GiB))"]
-
-[[benchmark]]
-model = "mlx-community/GLM-4.7-4bit"
-extra_constraints = ["All(Memory(>=256GiB))"]
-
-[[benchmark]]
-model = "mlx-community/MiniMax-M2.1-3bit"
-extra_constraints = ["All(Memory(>=256GiB))"]
-
-[[benchmark]]
-model = "mlx-community/MiniMax-M2.1-8bit"
-extra_constraints = ["All(Memory(>=256GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-235B-A22B-Instruct-2507-4bit"
-extra_constraints = ["All(Memory(>=256GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-Coder-Next-bf16"
-extra_constraints = ["All(Memory(>=256GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Step-3.5-Flash-4bit"
-extra_constraints = ["All(Memory(>=256GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Step-3.5-Flash-6bit"
-extra_constraints = ["All(Memory(>=256GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Step-3.5-Flash-8Bit"
-extra_constraints = ["All(Memory(>=256GiB))"]
-
-[[benchmark]]
-model = "mlx-community/DeepSeek-V3.1-4bit"
-extra_constraints = ["All(Memory(>=512GiB))"]
-
-[[benchmark]]
-model = "mlx-community/GLM-4.7-6bit"
-extra_constraints = ["All(Memory(>=512GiB))"]
-
-[[benchmark]]
-model = "mlx-community/GLM-4.7-8bit-gs32"
-extra_constraints = ["All(Memory(>=512GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-235B-A22B-Instruct-2507-8bit"
-extra_constraints = ["All(Memory(>=512GiB))"]
-
-[[benchmark]]
-model = "mlx-community/Qwen3-Coder-480B-A35B-Instruct-4bit"
-extra_constraints = ["All(Memory(>=512GiB))"]
--- a/dashboard/src/lib/components/ChatForm.svelte
+++ b/dashboard/src/lib/components/ChatForm.svelte
@@ -265,7 +265,6 @@

  function handleSubmit() {
    if ((!message.trim() && uploadedFiles.length === 0) || loading) return;
-    if (isEditOnlyWithoutImage) return;

    const content = message.trim();
    const files = [...uploadedFiles];
@@ -290,11 +289,7 @@
      if (imageFile.preview) {
        editImage(content, imageFile.preview);
      }
-    } else if (
-      currentModel &&
-      modelSupportsTextToImage(currentModel) &&
-      content
-    ) {
+    } else if (isImageModel() && content) {
      // Use image generation for text-to-image models
      generateImage(content);
    } else {
--- a/dashboard/src/lib/components/ChatMessages.svelte
+++ b/dashboard/src/lib/components/ChatMessages.svelte
@@ -9,7 +9,6 @@
    regenerateFromToken,
    setEditingImage,
  } from "$lib/stores/app.svelte";
-  import type { Message } from "$lib/stores/app.svelte";
  import type { MessageAttachment } from "$lib/stores/app.svelte";
  import MarkdownContent from "./MarkdownContent.svelte";
  import TokenHeatmap from "./TokenHeatmap.svelte";
@@ -225,7 +224,6 @@
  }

  function handleDeleteClick(messageId: string) {
-    if (loading) return;
    deleteConfirmId = messageId;
  }

@@ -256,7 +254,7 @@
 </script>

 <div class="flex flex-col gap-4 sm:gap-6 {className}">
-  {#each messageList as message, i (message.id)}
+  {#each messageList as message (message.id)}
    <div
      class="group flex {message.role === 'user'
        ? 'justify-end'
@@ -318,11 +316,9 @@
          <!-- Delete confirmation -->
          <div class="bg-red-500/10 border border-red-500/30 rounded-lg p-3">
            <p class="text-xs text-red-400 mb-3">
-              {#if i === messageList.length - 1}
-                Delete this message?
-              {:else}
-                Delete this message and all messages after it?
-              {/if}
+              Delete this message{message.role === "user"
+                ? " and all responses after it"
+                : ""}?
            </p>
            <div class="flex gap-2 justify-end">
              <button
@@ -754,13 +750,8 @@
            <!-- Delete button -->
            <button
              onclick={() => handleDeleteClick(message.id)}
-              disabled={loading}
-              class="p-1.5 transition-colors rounded {loading
-                ? 'text-exo-light-gray/30 cursor-not-allowed'
-                : 'text-exo-light-gray hover:text-red-400 hover:bg-red-500/10 cursor-pointer'}"
-              title={loading
-                ? "Cannot delete while generating"
-                : "Delete message"}
+              class="p-1.5 text-exo-light-gray hover:text-red-400 transition-colors rounded hover:bg-red-500/10 cursor-pointer"
+              title="Delete message"
            >
              <svg
                class="w-3.5 h-3.5"
--- a/dashboard/src/lib/components/PrefillProgressBar.svelte
+++ b/dashboard/src/lib/components/PrefillProgressBar.svelte
@@ -0,0 +1,51 @@
+<script lang="ts">
+  import type { PrefillProgress } from "$lib/stores/app.svelte";
+
+  interface Props {
+    progress: PrefillProgress;
+    class?: string;
+  }
+
+  let { progress, class: className = "" }: Props = $props();
+
+  const percentage = $derived(
+    progress.total > 0
+      ? Math.round((progress.processed / progress.total) * 100)
+      : 0,
+  );
+
+  function formatTokenCount(count: number): string {
+    if (count >= 1000) {
+      return `${(count / 1000).toFixed(1)}k`;
+    }
+    return count.toString();
+  }
+</script>
+
+<div class="prefill-progress {className}">
+  <div
+    class="flex items-center justify-between text-xs text-exo-light-gray mb-1"
+  >
+    <span>Processing prompt</span>
+    <span class="font-mono">
+      {formatTokenCount(progress.processed)} / {formatTokenCount(
+        progress.total,
+      )} tokens
+    </span>
+  </div>
+  <div class="h-1.5 bg-exo-black/60 rounded-full overflow-hidden">
+    <div
+      class="h-full bg-exo-yellow rounded-full transition-all duration-150 ease-out"
+      style="width: {percentage}%"
+    ></div>
+  </div>
+  <div class="text-right text-xs text-exo-light-gray/70 mt-0.5 font-mono">
+    {percentage}%
+  </div>
+</div>
+
+<style>
+  .prefill-progress {
+    width: 100%;
+  }
+</style>
--- a/dashboard/src/lib/stores/app.svelte.ts
+++ b/dashboard/src/lib/stores/app.svelte.ts
@@ -273,6 +273,11 @@ export interface TokenData {
  topLogprobs: TopLogprob[];
 }

+export interface PrefillProgress {
+  processed: number;
+  total: number;
+}
+
 export interface Message {
  id: string;
  role: "user" | "assistant" | "system";
--- a/dashboard/src/routes/downloads/+page.svelte
+++ b/dashboard/src/routes/downloads/+page.svelte
--- a/flake.nix
+++ b/flake.nix
@@ -115,7 +115,7 @@
          packages = lib.optionalAttrs pkgs.stdenv.hostPlatform.isDarwin (
            let
              uvLock = builtins.fromTOML (builtins.readFile ./uv.lock);
-              mlxPackage = builtins.head (builtins.filter (p: p.name == "mlx" && p.source ? git) uvLock.package);
+              mlxPackage = builtins.head (builtins.filter (p: p.name == "mlx") uvLock.package);
              uvLockMlxVersion = mlxPackage.version;
            in
            {
--- a/nix/mlx.nix
+++ b/nix/mlx.nix
@@ -41,16 +41,16 @@ let

  mlx = stdenv.mkDerivation rec {
    pname = "mlx";
-    version = let v = "0.30.7.dev20260217+50487b41"; in
+    version = let v = "0.30.6"; in
      assert v == uvLockMlxVersion || throw "MLX version mismatch: nix/mlx.nix has ${v} but uv.lock has ${uvLockMlxVersion}. Update both the version and hash in nix/mlx.nix.";
      v;
    pyproject = true;

    src = fetchFromGitHub {
-      owner = "rltakashige";
-      repo = "mlx-jaccl-fix-small-recv";
-      rev = "50487b4141f3c951122655db3b83df5146c1fbeb";
-      hash = "sha256-IL4a9vMX5nocgJU1WG4zE8hArHkHJtnh4sdYh3od5zU=";
+      owner = "ml-explore";
+      repo = "mlx";
+      tag = "v${version}";
+      hash = "sha256-avD5EGhwgmPdXLAyQSqTO6AXk/W3ziH+f6AetjK3Sdo=";
    };

    patches = [
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -17,9 +17,9 @@ dependencies = [
    "loguru>=0.7.3",
    "exo_pyo3_bindings", # rust bindings
    "anyio==4.11.0",
-    "mlx; sys_platform == 'darwin'",
+    "mlx==0.30.6; sys_platform == 'darwin'",
    "mlx[cpu]==0.30.6; sys_platform == 'linux'",
-    "mlx-lm==0.30.7",
+    "mlx-lm==0.30.6",
    "tiktoken>=0.12.0", # required for kimi k2 tokenizer
    "hypercorn>=0.18.0",
    "openai-harmony>=0.0.8",
@@ -64,7 +64,6 @@ members = [

 [tool.uv.sources]
 exo_pyo3_bindings = { workspace = true }
-mlx = { git = "https://github.com/rltakashige/mlx-jaccl-fix-small-recv.git", branch = "address-rdma-gpu-locks", marker = "sys_platform == 'darwin'" }
 #mlx-lm = { git = "https://github.com/davidmcc73/mlx-lm", branch = "stable" }
 # Uncomment to use local mlx/mlx-lm development versions:
 # mlx = { path = "/Users/Shared/mlx", editable=true }
@@ -133,7 +132,7 @@ markers = [
 env = [
  "EXO_TESTS=1"
 ]
-addopts = "-m 'not slow' --ignore=tests/start_distributed_test.py"
+addopts = "-m 'not slow'"
 filterwarnings = [
    "ignore:builtin type Swig:DeprecationWarning",
 ]
--- a/python/parts.nix
+++ b/python/parts.nix
@@ -14,9 +14,7 @@

      # Override overlay to inject Nix-built components
      exoOverlay = final: prev: {
-        # Replace workspace exo_pyo3_bindings with Nix-built wheel.
-        # Preserve passthru so mkVirtualEnv can resolve dependency groups.
-        # Copy .pyi stub + py.typed marker so basedpyright can find the types.
+        # Replace workspace exo_pyo3_bindings with Nix-built wheel
        exo-pyo3-bindings = pkgs.stdenv.mkDerivation {
          pname = "exo-pyo3-bindings";
          version = "0.1.0";
@@ -24,12 +22,6 @@
          # Install from pre-built wheel
          nativeBuildInputs = [ final.pyprojectWheelHook ];
          dontStrip = true;
-          passthru = prev.exo-pyo3-bindings.passthru or { };
-          postInstall = ''
-            local siteDir=$out/${final.python.sitePackages}/exo_pyo3_bindings
-            cp ${inputs.self}/rust/exo_pyo3_bindings/exo_pyo3_bindings.pyi $siteDir/
-            touch $siteDir/py.typed
-          '';
        };
      };

@@ -37,47 +29,17 @@

      # Overlay to provide build systems and custom packages
      buildSystemsOverlay = final: prev: {
+        # Use our pure Nix-built MLX with Metal support
+        mlx = self'.packages.mlx;
+
        # mlx-lm is a git dependency that needs setuptools
        mlx-lm = prev.mlx-lm.overrideAttrs (old: {
          nativeBuildInputs = (old.nativeBuildInputs or [ ]) ++ [
            final.setuptools
          ];
        });
-      } // lib.optionalAttrs pkgs.stdenv.hostPlatform.isDarwin {
-        # Use our pure Nix-built MLX with Metal support (macOS only)
-        mlx = self'.packages.mlx;
      };

-      # Additional overlay for Linux-specific fixes (type checking env).
-      # Native wheels have shared lib dependencies we don't need at type-check time.
-      linuxOverlay = final: prev:
-        let
-          ignoreMissing = drv: drv.overrideAttrs { autoPatchelfIgnoreMissingDeps = [ "*" ]; };
-          nvidiaPackages = lib.filterAttrs (name: _: lib.hasPrefix "nvidia-" name) prev;
-        in
-        lib.optionalAttrs pkgs.stdenv.hostPlatform.isLinux (
-          (lib.mapAttrs (_: ignoreMissing) nvidiaPackages) // {
-            mlx = ignoreMissing prev.mlx;
-            mlx-cuda-13 = prev.mlx-cuda-13.overrideAttrs (old: {
-              buildInputs = (old.buildInputs or [ ]) ++ [
-                final.nvidia-cublas
-                final.nvidia-cuda-nvrtc
-                final.nvidia-cudnn-cu13
-                final.nvidia-nccl-cu13
-              ];
-              preFixup = ''
-                addAutoPatchelfSearchPath ${final.nvidia-cublas}
-                addAutoPatchelfSearchPath ${final.nvidia-cuda-nvrtc}
-                addAutoPatchelfSearchPath ${final.nvidia-cudnn-cu13}
-                addAutoPatchelfSearchPath ${final.nvidia-nccl-cu13}
-              '';
-              autoPatchelfIgnoreMissingDeps = [ "libcuda.so.1" ];
-            });
-            torch = ignoreMissing prev.torch;
-            triton = ignoreMissing prev.triton;
-          }
-        );
-
      pythonSet = (pkgs.callPackage inputs.pyproject-nix.build.packages {
        inherit python;
      }).overrideScope (
@@ -86,28 +48,16 @@
          overlay
          exoOverlay
          buildSystemsOverlay
-          linuxOverlay
        ]
      );
-      # mlx-cpu and mlx-cuda-13 both ship mlx/ site-packages files; keep first.
-      # mlx-cpu/mlx-cuda-13 and nvidia-cudnn-cu12/cu13 ship overlapping files.
-      venvCollisionPaths = lib.optionals pkgs.stdenv.hostPlatform.isLinux [
-        "lib/python3.13/site-packages/mlx*"
-        "lib/python3.13/site-packages/nvidia*"
-      ];
-
-      exoVenv = (pythonSet.mkVirtualEnv "exo-env" workspace.deps.default).overrideAttrs {
-        venvIgnoreCollisions = venvCollisionPaths;
-      };
+      exoVenv = pythonSet.mkVirtualEnv "exo-env" workspace.deps.default;

      # Virtual environment with dev dependencies for testing
-      testVenv = (pythonSet.mkVirtualEnv "exo-test-env" (
+      testVenv = pythonSet.mkVirtualEnv "exo-test-env" (
        workspace.deps.default // {
          exo = [ "dev" ]; # Include pytest, pytest-asyncio, pytest-env
        }
-      )).overrideAttrs {
-        venvIgnoreCollisions = venvCollisionPaths;
-      };
+      );

      mkPythonScript = name: path: pkgs.writeShellApplication {
        inherit name;
@@ -168,21 +118,6 @@
          ${pkgs.ruff}/bin/ruff check ${inputs.self}
          touch $out
        '';
-
-        # Hermetic basedpyright type checking
-        typecheck = pkgs.runCommand "typecheck"
-          {
-            nativeBuildInputs = [
-              testVenv
-              pkgs.basedpyright
-            ];
-          }
-          ''
-            cd ${inputs.self}
-            export HOME=$TMPDIR
-            basedpyright --pythonpath ${testVenv}/bin/python
-            touch $out
-          '';
      };
    };
 }
--- a/rust/exo_pyo3_bindings/Cargo.toml
+++ b/rust/exo_pyo3_bindings/Cargo.toml
@@ -25,38 +25,53 @@ workspace = true
 networking = { workspace = true }

 # interop
-pyo3 = { version = "0.27.2", features = [
-    # "abi3-py313", # tells pyo3 (and maturin) to build using the stable ABI with minimum Python version 3.13
-    # "nightly", # enables better-supported GIL integration
+pyo3 = { version = "0.27.1", features = [
+    # "abi3-py311", # tells pyo3 (and maturin) to build using the stable ABI with minimum Python version 3.11
+    "nightly", # enables better-supported GIL integration
    "experimental-async", # async support in #[pyfunction] & #[pymethods]
    #"experimental-inspect", # inspection of generated binary => easier to automate type-hint generation
    #"py-clone", # adding Clone-ing of `Py<T>` without GIL (may cause panics - remove if panics happen)
-    # "multiple-pymethods", # allows multiple #[pymethods] sections per class
+    "multiple-pymethods", # allows multiple #[pymethods] sections per class

    # integrations with other libraries
-    # "arc_lock", "bigdecimal", "either", "hashbrown", "indexmap", "num-bigint", "num-complex", "num-rational",
-    # "ordered-float", "rust_decimal", "smallvec",
+    "arc_lock", "bigdecimal", "either", "hashbrown", "indexmap", "num-bigint", "num-complex", "num-rational",
+    "ordered-float", "rust_decimal", "smallvec",
    # "anyhow", "chrono", "chrono-local", "chrono-tz", "eyre", "jiff-02", "lock_api", "parking-lot", "time",  "serde",
 ] }
-pyo3-stub-gen = { version = "0.19.0" }
+pyo3-stub-gen = { version = "0.17.2" }
 pyo3-async-runtimes = { version = "0.27.0", features = ["attributes", "tokio-runtime", "testing"] }
 pyo3-log = "0.13.2"

 # macro dependencies
 extend = { workspace = true }
 delegate = { workspace = true }
+impl-trait-for-tuples = { workspace = true }
+derive_more = { workspace = true }
 pin-project = { workspace = true }

 # async runtime
 tokio = { workspace = true, features = ["full", "tracing"] }
-futures-lite = { workspace = true }
+futures = { workspace = true }

 # utility dependencies
+once_cell = "1.21.3"
+thread_local = "1.1.9"
 util = { workspace = true }
+thiserror = { workspace = true }
+#internment = { workspace = true }
+#recursion = { workspace = true }
+#generativity = { workspace = true }
+#itertools = { workspace = true }
+

 # Tracing
+#tracing = "0.1"
+#tracing-subscriber = "0.3"
+#console-subscriber = "0.1.5"
+#tracing-log = "0.2.0"
 log = { workspace = true }
 env_logger = "0.11"

+
 # Networking
 libp2p = { workspace = true, features = ["full"] }
--- a/rust/exo_pyo3_bindings/exo_pyo3_bindings.pyi
+++ b/rust/exo_pyo3_bindings/exo_pyo3_bindings.pyi
@@ -1,85 +1,155 @@
 # This file is automatically generated by pyo3_stub_gen
-# ruff: noqa: E501, F401, F403, F405
+# ruff: noqa: E501, F401

 import builtins
+import enum
 import typing
-__all__ = [
-    "AllQueuesFullError",
-    "Keypair",
-    "NoPeersSubscribedToTopicError",
-    "PyMessage",
-    "PySwarm",
-]

@typing.final
 class AllQueuesFullError(builtins.Exception):
-    def __new__(cls, *_a: typing.Any) -> AllQueuesFullError: ...
+    def __new__(cls, *args: typing.Any) -> AllQueuesFullError: ...
+    def __repr__(self) -> builtins.str: ...
    def __str__(self) -> builtins.str: ...

+@typing.final
+class ConnectionUpdate:
+    @property
+    def update_type(self) -> ConnectionUpdateType:
+        r"""
+        Whether this is a connection or disconnection event
+        """
+    @property
+    def peer_id(self) -> PeerId:
+        r"""
+        Identity of the peer that we have connected to or disconnected from.
+        """
+    @property
+    def remote_ipv4(self) -> builtins.str:
+        r"""
+        Remote connection's IPv4 address.
+        """
+    @property
+    def remote_tcp_port(self) -> builtins.int:
+        r"""
+        Remote connection's TCP port.
+        """
+
@typing.final
 class Keypair:
    r"""
    Identity keypair of a node.
    """
    @staticmethod
-    def generate() -> Keypair:
+    def generate_ed25519() -> Keypair:
        r"""
        Generate a new Ed25519 keypair.
        """
    @staticmethod
-    def deserialize(bytes: bytes) -> Keypair:
+    def generate_ecdsa() -> Keypair:
+        r"""
+        Generate a new ECDSA keypair.
+        """
+    @staticmethod
+    def generate_secp256k1() -> Keypair:
+        r"""
+        Generate a new Secp256k1 keypair.
+        """
+    @staticmethod
+    def from_protobuf_encoding(bytes: bytes) -> Keypair:
        r"""
        Decode a private key from a protobuf structure and parse it as a `Keypair`.
        """
-    def serialize(self) -> bytes:
+    @staticmethod
+    def rsa_from_pkcs8(bytes: bytes) -> Keypair:
+        r"""
+        Decode an keypair from a DER-encoded secret key in PKCS#8 `PrivateKeyInfo`
+        format (i.e. unencrypted) as defined in [RFC5208].
+        
+        [RFC5208]: https://tools.ietf.org/html/rfc5208#section-5
+        """
+    @staticmethod
+    def secp256k1_from_der(bytes: bytes) -> Keypair:
+        r"""
+        Decode a keypair from a DER-encoded Secp256k1 secret key in an `ECPrivateKey`
+        structure as defined in [RFC5915].
+        
+        [RFC5915]: https://tools.ietf.org/html/rfc5915
+        """
+    @staticmethod
+    def ed25519_from_bytes(bytes: bytes) -> Keypair: ...
+    def to_protobuf_encoding(self) -> bytes:
        r"""
        Encode a private key as protobuf structure.
        """
-    def to_string(self) -> builtins.str:
+    def to_peer_id(self) -> PeerId:
        r"""
        Convert the `Keypair` into the corresponding `PeerId`.
        """

@typing.final
-class NoPeersSubscribedToTopicError(builtins.Exception):
-    def __new__(cls, *_a: typing.Any) -> NoPeersSubscribedToTopicError: ...
-    def __str__(self) -> builtins.str: ...
-
-class PyMessage:
-    @typing.final
-    class Connection(PyMessage):
-        __match_args__ = ("node_id", "connected",)
-        @property
-        def node_id(self) -> builtins.str: ...
-        @property
-        def connected(self) -> builtins.bool: ...
-        def __new__(cls, node_id: builtins.str, connected: builtins.bool) -> PyMessage.Connection: ...
-    
-    @typing.final
-    class Gossip(PyMessage):
-        __match_args__ = ("node_id", "topic", "data",)
-        @property
-        def node_id(self) -> builtins.str: ...
-        @property
-        def topic(self) -> builtins.str: ...
-        @property
-        def data(self) -> bytes: ...
-        def __new__(cls, node_id: builtins.str, topic: builtins.str, data: bytes) -> PyMessage.Gossip: ...
-    
-    ...
+class Multiaddr:
+    r"""
+    Representation of a Multiaddr.
+    """
+    @staticmethod
+    def empty() -> Multiaddr:
+        r"""
+        Create a new, empty multiaddress.
+        """
+    @staticmethod
+    def with_capacity(n: builtins.int) -> Multiaddr:
+        r"""
+        Create a new, empty multiaddress with the given capacity.
+        """
+    @staticmethod
+    def from_bytes(bytes: bytes) -> Multiaddr:
+        r"""
+        Parse a `Multiaddr` value from its byte slice representation.
+        """
+    @staticmethod
+    def from_string(string: builtins.str) -> Multiaddr:
+        r"""
+        Parse a `Multiaddr` value from its string representation.
+        """
+    def len(self) -> builtins.int:
+        r"""
+        Return the length in bytes of this multiaddress.
+        """
+    def is_empty(self) -> builtins.bool:
+        r"""
+        Returns true if the length of this multiaddress is 0.
+        """
+    def to_bytes(self) -> bytes:
+        r"""
+        Return a copy of this [`Multiaddr`]'s byte representation.
+        """
+    def to_string(self) -> builtins.str:
+        r"""
+        Convert a Multiaddr to a string.
+        """

@typing.final
-class PySwarm:
-    def __new__(cls, identity: Keypair) -> PySwarm: ...
-    async def recv(self) -> PyMessage:
+class NetworkingHandle:
+    def __new__(cls, identity: Keypair) -> NetworkingHandle: ...
+    async def connection_update_recv(self) -> ConnectionUpdate:
        r"""
-        Receives the next message from networking.
+        Receives the next `ConnectionUpdate` from networking.
        """
-    async def gossipsub_subscribe(self, topic: builtins.str) -> None:
+    async def connection_update_recv_many(self, limit: builtins.int) -> builtins.list[ConnectionUpdate]:
+        r"""
+        Receives at most `limit` `ConnectionUpdate`s from networking and returns them.
+        
+        For `limit = 0`, an empty collection of `ConnectionUpdate`s will be returned immediately.
+        For `limit > 0`, if there are no `ConnectionUpdate`s in the channel's queue this method
+        will sleep until a `ConnectionUpdate`s is sent.
+        """
+    async def gossipsub_subscribe(self, topic: builtins.str) -> builtins.bool:
        r"""
        Subscribe to a `GossipSub` topic.
+        
+        Returns `True` if the subscription worked. Returns `False` if we were already subscribed.
        """
-    async def gossipsub_unsubscribe(self, topic: builtins.str) -> None:
+    async def gossipsub_unsubscribe(self, topic: builtins.str) -> builtins.bool:
        r"""
        Unsubscribes from a `GossipSub` topic.
        
@@ -87,6 +157,65 @@ class PySwarm:
        """
    async def gossipsub_publish(self, topic: builtins.str, data: bytes) -> None:
        r"""
-        Publishes a message to the network on a specific topic.
+        Publishes a message with multiple topics to the `GossipSub` network.
+        
+        If no peers are found that subscribe to this topic, throws `NoPeersSubscribedToTopicError` exception.
+        """
+    async def gossipsub_recv(self) -> tuple[builtins.str, bytes]:
+        r"""
+        Receives the next message from the `GossipSub` network.
+        """
+    async def gossipsub_recv_many(self, limit: builtins.int) -> builtins.list[tuple[builtins.str, bytes]]:
+        r"""
+        Receives at most `limit` messages from the `GossipSub` network and returns them.
+        
+        For `limit = 0`, an empty collection of messages will be returned immediately.
+        For `limit > 0`, if there are no messages in the channel's queue this method
+        will sleep until a message is sent.
        """

+@typing.final
+class NoPeersSubscribedToTopicError(builtins.Exception):
+    def __new__(cls, *args: typing.Any) -> NoPeersSubscribedToTopicError: ...
+    def __repr__(self) -> builtins.str: ...
+    def __str__(self) -> builtins.str: ...
+
+@typing.final
+class PeerId:
+    r"""
+    Identifier of a peer of the network.
+    
+    The data is a `CIDv0` compatible multihash of the protobuf encoded public key of the peer
+    as specified in [specs/peer-ids](https://github.com/libp2p/specs/blob/master/peer-ids/peer-ids.md).
+    """
+    @staticmethod
+    def random() -> PeerId:
+        r"""
+        Generates a random peer ID from a cryptographically secure PRNG.
+        
+        This is useful for randomly walking on a DHT, or for testing purposes.
+        """
+    @staticmethod
+    def from_bytes(bytes: bytes) -> PeerId:
+        r"""
+        Parses a `PeerId` from bytes.
+        """
+    def to_bytes(self) -> bytes:
+        r"""
+        Returns a raw bytes representation of this `PeerId`.
+        """
+    def to_base58(self) -> builtins.str:
+        r"""
+        Returns a base-58 encoded string of this `PeerId`.
+        """
+    def __repr__(self) -> builtins.str: ...
+    def __str__(self) -> builtins.str: ...
+
+@typing.final
+class ConnectionUpdateType(enum.Enum):
+    r"""
+    Connection or disconnection event discriminant type.
+    """
+    Connected = ...
+    Disconnected = ...
+
--- a/rust/exo_pyo3_bindings/src/allow_threading.rs
+++ b/rust/exo_pyo3_bindings/src/allow_threading.rs
@@ -1,4 +1,8 @@
-//! See: <https://pyo3.rs/v0.27.2/async-await.html#detaching-from-the-interpreter-across-await>
+//! SEE: https://pyo3.rs/v0.26.0/async-await.html#detaching-from-the-interpreter-across-await
+//!
+
+use pin_project::pin_project;
+use pyo3::marker::Ungil;
 use pyo3::prelude::*;
 use std::{
    future::Future,
@@ -6,17 +10,31 @@ use std::{
    task::{Context, Poll},
 };

-pub struct AllowThreads<F>(pub(crate) F);
+/// SEE: https://pyo3.rs/v0.26.0/async-await.html#detaching-from-the-interpreter-across-await
+#[pin_project]
+#[repr(transparent)]
+pub(crate) struct AllowThreads<F>(#[pin] F);
+
+impl<F> AllowThreads<F>
+where
+    Self: Future,
+{
+    pub fn new(f: F) -> Self {
+        Self(f)
+    }
+}

 impl<F> Future for AllowThreads<F>
 where
-    F: Future + Unpin + Send,
-    F::Output: Send,
+    F: Future + Ungil,
+    F::Output: Ungil,
 {
    type Output = F::Output;

-    fn poll(mut self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Self::Output> {
+    fn poll(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Self::Output> {
        let waker = cx.waker();
-        Python::attach(|py| py.detach(|| pin!(&mut self.0).poll(&mut Context::from_waker(waker))))
+        Python::with_gil(|py| {
+            py.allow_threads(|| self.project().0.poll(&mut Context::from_waker(waker)))
+        })
    }
 }
--- a/rust/exo_pyo3_bindings/src/examples/mod.rs
+++ b/rust/exo_pyo3_bindings/src/examples/mod.rs
@@ -0,0 +1,240 @@
+//! This module exists to hold examples of some pyo3 patterns that may be too complex to
+//! re-create from scratch, but too inhomogenous to create an abstraction/wrapper around.
+//!
+//! Pattern examples include:
+//!  - Async task handles: with GC-integrated cleanup
+//!  - Sync/async callbacks from python: with propper eventloop handling
+//!
+//! Mutability pattern: https://pyo3.rs/v0.26.0/async-await.html#send--static-constraint
+//!  - Store mutable fields in tokio's `Mutex<T>`
+//!  - For async code: take `&self` and `.lock().await`
+//!  - For sync code: take `&mut self` and `.get_mut()`
+
+use crate::ext::{PyResultExt as _, ResultExt as _, TokioRuntimeExt as _};
+use futures::FutureExt as _;
+use futures::future::BoxFuture;
+use pyo3::exceptions::PyRuntimeError;
+use pyo3::prelude::{PyModule, PyModuleMethods as _};
+use pyo3::{
+    Bound, Py, PyAny, PyErr, PyResult, PyTraverseError, PyVisit, Python, pyclass, pymethods,
+};
+use std::time::Duration;
+use tokio::sync::mpsc;
+use tokio::sync::mpsc::error::TryRecvError;
+
+fn needs_tokio_runtime() {
+    tokio::runtime::Handle::current();
+}
+
+type SyncCallback = Box<dyn Fn() + Send + Sync>;
+type AsyncCallback = Box<dyn Fn() -> BoxFuture<'static, ()> + Send + Sync>;
+
+enum AsyncTaskMessage {
+    SyncCallback(SyncCallback),
+    AsyncCallback(AsyncCallback),
+}
+
+async fn async_task(
+    sender: mpsc::UnboundedSender<()>,
+    mut receiver: mpsc::UnboundedReceiver<AsyncTaskMessage>,
+) {
+    log::info!("RUST: async task started");
+
+    // task state
+    let mut interval = tokio::time::interval(Duration::from_secs(1));
+
+    let mut sync_cbs: Vec<SyncCallback> = vec![];
+    let mut async_cbs: Vec<AsyncCallback> = vec![];
+
+    loop {
+        tokio::select! {
+            // handle incoming messages from task-handle
+            message = receiver.recv() => {
+                // handle closed channel by exiting
+                let Some(message) = message else {
+                    log::info!("RUST: channel closed");
+                    break;
+                };
+
+                // dispatch incoming event
+                match message {
+                    AsyncTaskMessage::SyncCallback(cb) => {
+                        sync_cbs.push(cb);
+                    }
+                    AsyncTaskMessage::AsyncCallback(cb) => {
+                        async_cbs.push(cb);
+                    }
+                }
+            }
+
+            // handle all other events
+            _ = interval.tick() => {
+                log::info!("RUST: async task tick");
+
+                // call back all sync callbacks
+                for cb in &sync_cbs {
+                    cb();
+                }
+
+                // call back all async callbacks
+                for cb in &async_cbs {
+                    cb().await;
+                }
+
+                // send event on unbounded channel
+                sender.send(()).expect("handle receiver cannot be closed/dropped");
+            }
+        }
+    }
+
+    log::info!("RUST: async task stopped");
+}
+
+// #[gen_stub_pyclass]
+#[pyclass(name = "AsyncTaskHandle")]
+#[derive(Debug)]
+struct PyAsyncTaskHandle {
+    sender: Option<mpsc::UnboundedSender<AsyncTaskMessage>>,
+    receiver: mpsc::UnboundedReceiver<()>,
+}
+
+#[allow(clippy::expect_used)]
+impl PyAsyncTaskHandle {
+    const fn sender(&self) -> &mpsc::UnboundedSender<AsyncTaskMessage> {
+        self.sender
+            .as_ref()
+            .expect("The sender should only be None after de-initialization.")
+    }
+
+    const fn sender_mut(&mut self) -> &mpsc::UnboundedSender<AsyncTaskMessage> {
+        self.sender
+            .as_mut()
+            .expect("The sender should only be None after de-initialization.")
+    }
+
+    const fn new(
+        sender: mpsc::UnboundedSender<AsyncTaskMessage>,
+        receiver: mpsc::UnboundedReceiver<()>,
+    ) -> Self {
+        Self {
+            sender: Some(sender),
+            receiver,
+        }
+    }
+}
+
+// #[gen_stub_pymethods]
+#[pymethods]
+impl PyAsyncTaskHandle {
+    #[new]
+    fn py_new(py: Python<'_>) -> PyResult<Self> {
+        use pyo3_async_runtimes::tokio::get_runtime;
+
+        // create communication channel TOWARDS our task
+        let (h_sender, t_receiver) = mpsc::unbounded_channel::<AsyncTaskMessage>();
+
+        // create communication channel FROM our task
+        let (t_sender, h_receiver) = mpsc::unbounded_channel::<()>();
+
+        // perform necessary setup within tokio context - or it crashes
+        let () = get_runtime().block_on(async { needs_tokio_runtime() });
+
+        // spawn tokio task with this thread's task-locals - without this, async callbacks on the new threads will not work!!
+        _ = get_runtime().spawn_with_scope(py, async move {
+            async_task(t_sender, t_receiver).await;
+        });
+        Ok(Self::new(h_sender, h_receiver))
+    }
+
+    /// NOTE: exceptions in callbacks are silently ignored until end of execution
+    fn add_sync_callback(
+        &self,
+        // #[gen_stub(override_type(
+        //     type_repr="collections.abc.Callable[[], None]",
+        //     imports=("collections.abc")
+        // ))]
+        callback: Py<PyAny>,
+    ) -> PyResult<()> {
+        // blocking call to async method -> can do non-blocking if needed
+        self.sender()
+            .send(AsyncTaskMessage::SyncCallback(Box::new(move || {
+                _ = Python::with_gil(|py| callback.call0(py).write_unraisable_with(py));
+            })))
+            .pyerr()?;
+        Ok(())
+    }
+
+    /// NOTE: exceptions in callbacks are silently ignored until end of execution
+    fn add_async_callback(
+        &self,
+        // #[gen_stub(override_type(
+        //     type_repr="collections.abc.Callable[[], collections.abc.Awaitable[None]]",
+        //     imports=("collections.abc")
+        // ))]
+        callback: Py<PyAny>,
+    ) -> PyResult<()> {
+        // blocking call to async method -> can do non-blocking if needed
+        self.sender()
+            .send(AsyncTaskMessage::AsyncCallback(Box::new(move || {
+                let c = Python::with_gil(|py| callback.clone_ref(py));
+                async move {
+                    if let Some(f) = Python::with_gil(|py| {
+                        let coroutine = c.call0(py).write_unraisable_with(py)?;
+                        pyo3_async_runtimes::tokio::into_future(coroutine.into_bound(py))
+                            .write_unraisable_with(py)
+                    }) {
+                        _ = f.await.write_unraisable();
+                    }
+                }
+                .boxed()
+            })))
+            .pyerr()?;
+        Ok(())
+    }
+
+    async fn receive_unit(&mut self) -> PyResult<()> {
+        self.receiver
+            .recv()
+            .await
+            .ok_or(PyErr::new::<PyRuntimeError, _>(
+                "cannot receive unit on closed channel",
+            ))
+    }
+
+    fn drain_units(&mut self) -> PyResult<i32> {
+        let mut cnt = 0;
+        loop {
+            match self.receiver.try_recv() {
+                Err(TryRecvError::Disconnected) => {
+                    return Err(PyErr::new::<PyRuntimeError, _>(
+                        "cannot receive unit on closed channel",
+                    ));
+                }
+                Err(TryRecvError::Empty) => return Ok(cnt),
+                Ok(()) => {
+                    cnt += 1;
+                    continue;
+                }
+            }
+        }
+    }
+
+    // #[gen_stub(skip)]
+    const fn __traverse__(&self, _visit: PyVisit<'_>) -> Result<(), PyTraverseError> {
+        Ok(()) // This is needed purely so `__clear__` can work
+    }
+
+    // #[gen_stub(skip)]
+    fn __clear__(&mut self) {
+        // TODO: may or may not need to await a "kill-signal" oneshot channel message,
+        //       to ensure that the networking task is done BEFORE exiting the clear function...
+        //       but this may require GIL?? and it may not be safe to call GIL here??
+        self.sender = None; // Using Option<T> as a trick to force `sender` channel to be dropped
+    }
+}
+
+pub fn examples_submodule(m: &Bound<'_, PyModule>) -> PyResult<()> {
+    m.add_class::<PyAsyncTaskHandle>()?;
+
+    Ok(())
+}
--- a/rust/exo_pyo3_bindings/src/ident.rs
+++ b/rust/exo_pyo3_bindings/src/ident.rs
@@ -1,47 +0,0 @@
-use crate::ext::ResultExt as _;
-use libp2p::identity::Keypair;
-use pyo3::prelude::{PyBytesMethods as _, PyModule, PyModuleMethods as _};
-use pyo3::types::PyBytes;
-use pyo3::{Bound, PyResult, Python, pyclass, pymethods};
-use pyo3_stub_gen::derive::{gen_stub_pyclass, gen_stub_pymethods};
-
-/// Identity keypair of a node.
-#[gen_stub_pyclass]
-#[pyclass(name = "Keypair", frozen)]
-#[repr(transparent)]
-pub struct PyKeypair(pub Keypair);
-
-#[gen_stub_pymethods]
-#[pymethods]
-#[allow(clippy::needless_pass_by_value)]
-impl PyKeypair {
-    /// Generate a new Ed25519 keypair.
-    #[staticmethod]
-    fn generate() -> Self {
-        Self(Keypair::generate_ed25519())
-    }
-
-    /// Decode a private key from a protobuf structure and parse it as a `Keypair`.
-    #[staticmethod]
-    fn deserialize(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
-        let bytes = Vec::from(bytes.as_bytes());
-        Ok(Self(Keypair::from_protobuf_encoding(&bytes).pyerr()?))
-    }
-
-    /// Encode a private key as protobuf structure.
-    fn serialize<'py>(&self, py: Python<'py>) -> PyResult<Bound<'py, PyBytes>> {
-        let bytes = self.0.to_protobuf_encoding().pyerr()?;
-        Ok(PyBytes::new(py, &bytes))
-    }
-
-    /// Convert the `Keypair` into the corresponding `PeerId`.
-    fn to_string(&self) -> String {
-        self.0.public().to_peer_id().to_base58()
-    }
-}
-
-pub fn ident_submodule(m: &Bound<'_, PyModule>) -> PyResult<()> {
-    m.add_class::<PyKeypair>()?;
-
-    Ok(())
-}
--- a/rust/exo_pyo3_bindings/src/lib.rs
+++ b/rust/exo_pyo3_bindings/src/lib.rs
@@ -4,13 +4,28 @@
 //!
 //!

-mod allow_threading;
-mod ident;
-mod networking;
+// enable Rust-unstable features for convenience
+#![feature(trait_alias)]
+#![feature(tuple_trait)]
+#![feature(unboxed_closures)]
+// #![feature(stmt_expr_attributes)]
+// #![feature(assert_matches)]
+// #![feature(async_fn_in_dyn_trait)]
+// #![feature(async_for_loop)]
+// #![feature(auto_traits)]
+// #![feature(negative_impls)]
+
+extern crate core;
+mod allow_threading;
+mod examples;
+pub(crate) mod networking;
+pub(crate) mod pylibp2p;

-use crate::ident::ident_submodule;
 use crate::networking::networking_submodule;
+use crate::pylibp2p::ident::ident_submodule;
+use crate::pylibp2p::multiaddr::multiaddr_submodule;
 use pyo3::prelude::PyModule;
+use pyo3::prelude::*;
 use pyo3::{Bound, PyResult, pyclass, pymodule};
 use pyo3_stub_gen::define_stub_info_gatherer;

@@ -19,18 +34,35 @@ pub(crate) mod r#const {
    pub const MPSC_CHANNEL_SIZE: usize = 1024;
 }

+/// Namespace for all the type/trait aliases used by this crate.
+pub(crate) mod alias {
+    use std::error::Error;
+    use std::marker::Tuple;
+
+    pub trait SendFn<Args: Tuple + Send + 'static, Output> =
+        Fn<Args, Output = Output> + Send + 'static;
+
+    pub type AnyError = Box<dyn Error + Send + Sync + 'static>;
+    pub type AnyResult<T> = Result<T, AnyError>;
+}
+
 /// Namespace for crate-wide extension traits/methods
 pub(crate) mod ext {
    use crate::allow_threading::AllowThreads;
    use extend::ext;
-    use pyo3::exceptions::PyRuntimeError;
+    use pyo3::exceptions::{PyConnectionError, PyRuntimeError};
+    use pyo3::marker::Ungil;
    use pyo3::types::PyBytes;
-    use pyo3::{Py, PyResult, Python};
+    use pyo3::{Py, PyErr, PyResult, Python};
+    use tokio::runtime::Runtime;
+    use tokio::sync::mpsc;
+    use tokio::sync::mpsc::error::TryRecvError;
+    use tokio::task::JoinHandle;

    #[ext(pub, name = ByteArrayExt)]
    impl [u8] {
        fn pybytes(&self) -> Py<PyBytes> {
-            Python::attach(|py| PyBytes::new(py, self).unbind())
+            Python::with_gil(|py| PyBytes::new(py, self).unbind())
        }
    }

@@ -45,16 +77,120 @@ pub(crate) mod ext {
    }

    pub trait FutureExt: Future + Sized {
-        /// SEE: https://pyo3.rs/v0.27.2/async-await.html#detaching-from-the-interpreter-across-await
+        /// SEE: https://pyo3.rs/v0.26.0/async-await.html#detaching-from-the-interpreter-across-await
        fn allow_threads_py(self) -> AllowThreads<Self>
        where
            AllowThreads<Self>: Future,
        {
-            AllowThreads(self)
+            AllowThreads::new(self)
        }
    }

    impl<T: Future> FutureExt for T {}
+
+    #[ext(pub, name = PyErrExt)]
+    impl PyErr {
+        fn receiver_channel_closed() -> Self {
+            PyConnectionError::new_err("Receiver channel closed unexpectedly")
+        }
+    }
+
+    #[ext(pub, name = PyResultExt)]
+    impl<T> PyResult<T> {
+        fn write_unraisable(self) -> Option<T> {
+            Python::with_gil(|py| self.write_unraisable_with(py))
+        }
+
+        fn write_unraisable_with(self, py: Python<'_>) -> Option<T> {
+            match self {
+                Ok(v) => Some(v),
+                Err(e) => {
+                    // write error back to python
+                    e.write_unraisable(py, None);
+                    None
+                }
+            }
+        }
+    }
+
+    #[ext(pub, name = TokioRuntimeExt)]
+    impl Runtime {
+        fn spawn_with_scope<F>(&self, py: Python<'_>, future: F) -> PyResult<JoinHandle<F::Output>>
+        where
+            F: Future + Send + 'static,
+            F::Output: Send + 'static,
+        {
+            let locals = pyo3_async_runtimes::tokio::get_current_locals(py)?;
+            Ok(self.spawn(pyo3_async_runtimes::tokio::scope(locals, future)))
+        }
+    }
+
+    #[ext(pub, name = TokioMpscSenderExt)]
+    impl<T> mpsc::Sender<T> {
+        /// Sends a value, waiting until there is capacity.
+        ///
+        /// A successful send occurs when it is determined that the other end of the
+        /// channel has not hung up already. An unsuccessful send would be one where
+        /// the corresponding receiver has already been closed.
+        async fn send_py(&self, value: T) -> PyResult<()> {
+            self.send(value)
+                .await
+                .map_err(|_| PyErr::receiver_channel_closed())
+        }
+    }
+
+    #[ext(pub, name = TokioMpscReceiverExt)]
+    impl<T> mpsc::Receiver<T> {
+        /// Receives the next value for this receiver.
+        async fn recv_py(&mut self) -> PyResult<T> {
+            self.recv().await.ok_or_else(PyErr::receiver_channel_closed)
+        }
+
+        /// Receives at most `limit` values for this receiver and returns them.
+        ///
+        /// For `limit = 0`, an empty collection of messages will be returned immediately.
+        /// For `limit > 0`, if there are no messages in the channel's queue this method
+        /// will sleep until a message is sent.
+        async fn recv_many_py(&mut self, limit: usize) -> PyResult<Vec<T>> {
+            // get updates from receiver channel
+            let mut updates = Vec::with_capacity(limit);
+            let received = self.recv_many(&mut updates, limit).await;
+
+            // if we received zero items, then the channel was unexpectedly closed
+            if limit != 0 && received == 0 {
+                return Err(PyErr::receiver_channel_closed());
+            }
+
+            Ok(updates)
+        }
+
+        /// Tries to receive the next value for this receiver.
+        fn try_recv_py(&mut self) -> PyResult<Option<T>> {
+            match self.try_recv() {
+                Ok(v) => Ok(Some(v)),
+                Err(TryRecvError::Empty) => Ok(None),
+                Err(TryRecvError::Disconnected) => Err(PyErr::receiver_channel_closed()),
+            }
+        }
+    }
+}
+
+pub(crate) mod private {
+    use std::marker::Sized;
+
+    /// Sealed traits support
+    pub trait Sealed {}
+    impl<T: ?Sized> Sealed for T {}
+}
+
+/// A wrapper around [`Py`] that implements [`Clone`] using [`Python::with_gil`].
+#[repr(transparent)]
+pub(crate) struct ClonePy<T>(pub Py<T>);
+
+impl<T> Clone for ClonePy<T> {
+    fn clone(&self) -> Self {
+        Python::with_gil(|py| Self(self.0.clone_ref(py)))
+    }
 }

 /// A Python module implemented in Rust. The name of this function must match
@@ -65,9 +201,16 @@ fn main_module(m: &Bound<'_, PyModule>) -> PyResult<()> {
    // install logger
    pyo3_log::init();

+    // TODO: for now this is all NOT a submodule, but figure out how to make the submodule system
+    //       work with maturin, where the types generate correctly, in the right folder, without
+    //       too many importing issues...
    ident_submodule(m)?;
+    multiaddr_submodule(m)?;
    networking_submodule(m)?;

+    // top-level constructs
+    // TODO: ...
+
    Ok(())
 }

--- a/rust/exo_pyo3_bindings/src/networking.rs
+++ b/rust/exo_pyo3_bindings/src/networking.rs
@@ -1,24 +1,31 @@
+#![allow(
+    clippy::multiple_inherent_impl,
+    clippy::unnecessary_wraps,
+    clippy::unused_self,
+    clippy::needless_pass_by_value
+)]
+
 use crate::r#const::MPSC_CHANNEL_SIZE;
-use crate::ext::ResultExt as _;
-use crate::ext::{ByteArrayExt as _, FutureExt as _};
-use crate::ident::PyKeypair;
-use crate::networking::exception::{PyAllQueuesFullError, PyNoPeersSubscribedToTopicError};
+use crate::ext::{ByteArrayExt as _, FutureExt, PyErrExt as _};
+use crate::ext::{ResultExt as _, TokioMpscReceiverExt as _, TokioMpscSenderExt as _};
 use crate::pyclass;
-use futures_lite::FutureExt as _;
-use networking::swarm::{FromSwarm, Swarm, ToSwarm};
-use pyo3::coroutine::CancelHandle;
-use pyo3::exceptions::{PyConnectionError, PyRuntimeError};
-use pyo3::prelude::*;
+use crate::pylibp2p::ident::{PyKeypair, PyPeerId};
+use libp2p::futures::StreamExt as _;
+use libp2p::gossipsub::{IdentTopic, Message, MessageId, PublishError};
+use libp2p::swarm::SwarmEvent;
+use libp2p::{gossipsub, mdns};
+use networking::discovery;
+use networking::swarm::create_swarm;
+use pyo3::prelude::{PyModule, PyModuleMethods as _};
 use pyo3::types::PyBytes;
-use pyo3_async_runtimes::tokio::get_runtime;
-use pyo3_stub_gen::derive::{gen_stub_pyclass, gen_stub_pyclass_complex_enum, gen_stub_pymethods};
-use std::pin::pin;
-use std::sync::Arc;
-use tokio::sync::{Mutex, mpsc};
+use pyo3::{Bound, Py, PyErr, PyResult, PyTraverseError, PyVisit, Python, pymethods};
+use pyo3_stub_gen::derive::{gen_stub_pyclass, gen_stub_pyclass_enum, gen_stub_pymethods};
+use std::net::IpAddr;
+use tokio::sync::{Mutex, mpsc, oneshot};

 mod exception {
    use pyo3::types::PyTuple;
-    use pyo3::{exceptions::PyException, prelude::*};
+    use pyo3::{PyErrArguments, exceptions::PyException, prelude::*};
    use pyo3_stub_gen::derive::*;

    #[gen_stub_pyclass]
@@ -42,11 +49,16 @@ mod exception {
    #[pymethods]
    impl PyNoPeersSubscribedToTopicError {
        #[new]
-        #[pyo3(signature = (*_a))]
-        pub(crate) fn new(_a: &Bound<'_, PyTuple>) -> Self {
+        #[pyo3(signature = (*args))]
+        #[allow(unused_variables)]
+        pub(crate) fn new(args: &Bound<'_, PyTuple>) -> Self {
            Self {}
        }

+        fn __repr__(&self) -> String {
+            format!("PeerId(\"{}\")", Self::MSG)
+        }
+
        fn __str__(&self) -> String {
            Self::MSG.to_string()
        }
@@ -72,179 +84,489 @@ mod exception {
    #[pymethods]
    impl PyAllQueuesFullError {
        #[new]
-        #[pyo3(signature = (*_a))]
-        pub(crate) fn new(_a: &Bound<'_, PyTuple>) -> Self {
+        #[pyo3(signature = (*args))]
+        #[allow(unused_variables)]
+        pub(crate) fn new(args: &Bound<'_, PyTuple>) -> Self {
            Self {}
        }

+        fn __repr__(&self) -> String {
+            format!("PeerId(\"{}\")", Self::MSG)
+        }
+
        fn __str__(&self) -> String {
            Self::MSG.to_string()
        }
    }
 }

-#[gen_stub_pyclass]
-#[pyclass]
-struct PySwarm {
-    swarm: Arc<Mutex<Swarm>>,
-    from_swarm: Mutex<mpsc::Receiver<FromSwarm>>,
-    to_swarm: Mutex<mpsc::Sender<ToSwarm>>,
+/// Connection or disconnection event discriminant type.
+#[gen_stub_pyclass_enum]
+#[pyclass(eq, eq_int, name = "ConnectionUpdateType")]
+#[derive(Debug, Clone, PartialEq)]
+enum PyConnectionUpdateType {
+    Connected = 0,
+    Disconnected,
 }

-#[gen_stub_pyclass_complex_enum]
-#[pyclass]
-pub enum PyMessage {
-    Connection {
-        node_id: String,
-        connected: bool,
-    },
-    Gossip {
-        node_id: String,
+#[gen_stub_pyclass]
+#[pyclass(frozen, name = "ConnectionUpdate")]
+#[derive(Debug, Clone)]
+struct PyConnectionUpdate {
+    /// Whether this is a connection or disconnection event
+    #[pyo3(get)]
+    update_type: PyConnectionUpdateType,
+
+    /// Identity of the peer that we have connected to or disconnected from.
+    #[pyo3(get)]
+    peer_id: PyPeerId,
+
+    /// Remote connection's IPv4 address.
+    #[pyo3(get)]
+    remote_ipv4: String,
+
+    /// Remote connection's TCP port.
+    #[pyo3(get)]
+    remote_tcp_port: u16,
+}
+
+enum ToTask {
+    GossipsubSubscribe {
        topic: String,
-        data: Py<PyBytes>,
+        result_tx: oneshot::Sender<PyResult<bool>>,
+    },
+    GossipsubUnsubscribe {
+        topic: String,
+        result_tx: oneshot::Sender<bool>,
+    },
+    GossipsubPublish {
+        topic: String,
+        data: Vec<u8>,
+        result_tx: oneshot::Sender<PyResult<MessageId>>,
    },
 }
-impl TryFrom<FromSwarm> for PyMessage {
-    type Error = PyErr;
-    fn try_from(value: FromSwarm) -> Result<Self, Self::Error> {
-        match value {
-            FromSwarm::Discovered(nid) => Ok(PyMessage::Connection {
-                node_id: nid.to_base58(),
-                connected: true,
-            }),
-            FromSwarm::Expired(nid) => Ok(PyMessage::Connection {
-                node_id: nid.to_base58(),
-                connected: false,
-            }),
-            FromSwarm::Message(nid, topic, data) => Ok(PyMessage::Gossip {
-                node_id: nid.to_base58(),
-                topic,
-                data: data.pybytes(),
-            }),
-            FromSwarm::PublishError(e) => match e {
-                libp2p::gossipsub::PublishError::NoPeersSubscribedToTopic => {
-                    Err(PyNoPeersSubscribedToTopicError::new_err())
+
+#[allow(clippy::enum_glob_use)]
+async fn networking_task(
+    mut swarm: networking::swarm::Swarm,
+    mut to_task_rx: mpsc::Receiver<ToTask>,
+    connection_update_tx: mpsc::Sender<PyConnectionUpdate>,
+    gossipsub_message_tx: mpsc::Sender<(String, Vec<u8>)>,
+) {
+    use SwarmEvent::*;
+    use ToTask::*;
+    use mdns::Event::*;
+    use networking::swarm::BehaviourEvent::*;
+
+    log::info!("RUST: networking task started");
+
+    loop {
+        tokio::select! {
+            message = to_task_rx.recv() => {
+                // handle closed channel
+                let Some(message) = message else {
+                    log::info!("RUST: channel closed");
+                    break;
+                };
+
+                // dispatch incoming messages
+                match message {
+                    GossipsubSubscribe { topic, result_tx } => {
+                        // try to subscribe
+                        let result = swarm.behaviour_mut()
+                            .gossipsub.subscribe(&IdentTopic::new(topic));
+
+                        // send response oneshot
+                        if let Err(e) = result_tx.send(result.pyerr()) {
+                            log::error!("RUST: could not subscribe to gossipsub topic since channel already closed: {e:?}");
+                            continue;
+                        }
+                    }
+                    GossipsubUnsubscribe { topic, result_tx } => {
+                        // try to unsubscribe from the topic
+                        let result = swarm.behaviour_mut()
+                            .gossipsub.unsubscribe(&IdentTopic::new(topic));
+
+                        // send response oneshot (or exit if connection closed)
+                        if let Err(e) = result_tx.send(result) {
+                            log::error!("RUST: could not unsubscribe from gossipsub topic since channel already closed: {e:?}");
+                            continue;
+                        }
+                    }
+                    GossipsubPublish { topic, data, result_tx } => {
+                        // try to publish the data -> catch NoPeersSubscribedToTopic error & convert to correct exception
+                        let result = swarm.behaviour_mut().gossipsub.publish(
+                            IdentTopic::new(topic), data);
+                        let pyresult: PyResult<MessageId> = if let Err(PublishError::NoPeersSubscribedToTopic) = result {
+                            Err(exception::PyNoPeersSubscribedToTopicError::new_err())
+                        } else if let Err(PublishError::AllQueuesFull(_)) = result {
+                            Err(exception::PyAllQueuesFullError::new_err())
+                        } else {
+                            result.pyerr()
+                        };
+
+                        // send response oneshot (or exit if connection closed)
+                        if let Err(e) = result_tx.send(pyresult) {
+                            log::error!("RUST: could not publish gossipsub message since channel already closed: {e:?}");
+                            continue;
+                        }
+                    }
                }
-                libp2p::gossipsub::PublishError::AllQueuesFull(_) => {
-                    Err(PyAllQueuesFullError::new_err())
+            }
+
+            // architectural solution to this problem:
+            // create keep_alive behavior who's job it is to dial peers discovered by mDNS (and drop when expired)
+            //   -> it will emmit TRUE connected/disconnected events consumable elsewhere
+            //
+            // gossipsub will feed off-of dial attempts created by networking, and that will bootstrap its' peers list
+            // then for actual communication it will dial those peers if need-be
+            swarm_event = swarm.select_next_some() => {
+                match swarm_event {
+                    Behaviour(Gossipsub(gossipsub::Event::Message {
+                        message: Message {
+                            topic,
+                            data,
+                            ..
+                        },
+                        ..
+                    })) => {
+                        // topic-ID is just the topic hash!!! (since we used identity hasher)
+                        let message = (topic.into_string(), data);
+
+                        // send incoming message to channel (or exit if connection closed)
+                        if let Err(e) = gossipsub_message_tx.send(message).await {
+                            log::error!("RUST: could not send incoming gossipsub message since channel already closed: {e}");
+                            continue;
+                        }
+                    },
+                    Behaviour(Discovery(discovery::Event::ConnectionEstablished { peer_id, remote_ip, remote_tcp_port, .. })) => {
+                        // grab IPv4 string
+                        let remote_ipv4 = match remote_ip {
+                            IpAddr::V4(ip) => ip.to_string(),
+                            IpAddr::V6(ip) => {
+                                log::warn!("RUST: ignoring connection to IPv6 address: {ip}");
+                                continue;
+                            }
+                        };
+
+                        // send connection event to channel (or exit if connection closed)
+                        if let Err(e) = connection_update_tx.send(PyConnectionUpdate {
+                            update_type: PyConnectionUpdateType::Connected,
+                            peer_id: PyPeerId(peer_id),
+                            remote_ipv4,
+                            remote_tcp_port,
+                        }).await {
+                            log::error!("RUST: could not send connection update since channel already closed: {e}");
+                            continue;
+                        }
+                    },
+                    Behaviour(Discovery(discovery::Event::ConnectionClosed { peer_id, remote_ip, remote_tcp_port, .. })) => {
+                        // grab IPv4 string
+                        let remote_ipv4 = match remote_ip {
+                            IpAddr::V4(ip) => ip.to_string(),
+                            IpAddr::V6(ip) => {
+                                log::warn!("RUST: ignoring disconnection from IPv6 address: {ip}");
+                                continue;
+                            }
+                        };
+
+                        // send disconnection event to channel (or exit if connection closed)
+                        if let Err(e) = connection_update_tx.send(PyConnectionUpdate {
+                            update_type: PyConnectionUpdateType::Disconnected,
+                            peer_id: PyPeerId(peer_id),
+                            remote_ipv4,
+                            remote_tcp_port,
+                        }).await {
+                            log::error!("RUST: could not send connection update since channel already closed: {e}");
+                            continue;
+                        }
+                    },
+                    e => {
+                        log::info!("RUST: other event {e:?}");
+                    }
                }
-                e => Err(PyRuntimeError::new_err(e.to_string())),
-            },
+            }
        }
    }
+
+    log::info!("RUST: networking task stopped");
+}
+
+#[gen_stub_pyclass]
+#[pyclass(name = "NetworkingHandle")]
+#[derive(Debug)]
+struct PyNetworkingHandle {
+    // channels
+    to_task_tx: Option<mpsc::Sender<ToTask>>,
+    connection_update_rx: Mutex<mpsc::Receiver<PyConnectionUpdate>>,
+    gossipsub_message_rx: Mutex<mpsc::Receiver<(String, Vec<u8>)>>,
+}
+
+impl Drop for PyNetworkingHandle {
+    fn drop(&mut self) {
+        // TODO: may or may not need to await a "kill-signal" oneshot channel message,
+        //       to ensure that the networking task is done BEFORE exiting the clear function...
+        //       but this may require GIL?? and it may not be safe to call GIL here??
+        self.to_task_tx = None; // Using Option<T> as a trick to force channel to be dropped
+    }
+}
+
+#[allow(clippy::expect_used)]
+impl PyNetworkingHandle {
+    fn new(
+        to_task_tx: mpsc::Sender<ToTask>,
+        connection_update_rx: mpsc::Receiver<PyConnectionUpdate>,
+        gossipsub_message_rx: mpsc::Receiver<(String, Vec<u8>)>,
+    ) -> Self {
+        Self {
+            to_task_tx: Some(to_task_tx),
+            connection_update_rx: Mutex::new(connection_update_rx),
+            gossipsub_message_rx: Mutex::new(gossipsub_message_rx),
+        }
+    }
+
+    const fn to_task_tx(&self) -> &mpsc::Sender<ToTask> {
+        self.to_task_tx
+            .as_ref()
+            .expect("The sender should only be None after de-initialization.")
+    }
 }

 #[gen_stub_pymethods]
 #[pymethods]
-impl PySwarm {
+impl PyNetworkingHandle {
+    // NOTE: `async fn`s here that use `.await` will wrap the future in `.allow_threads_py()`
+    //       immediately beforehand to release the interpreter.
+    //       SEE: https://pyo3.rs/v0.26.0/async-await.html#detaching-from-the-interpreter-across-await
+
+    // ---- Lifecycle management methods ----
+
    #[new]
    fn py_new(identity: Bound<'_, PyKeypair>) -> PyResult<Self> {
        use pyo3_async_runtimes::tokio::get_runtime;

+        // create communication channels
+        let (to_task_tx, to_task_rx) = mpsc::channel(MPSC_CHANNEL_SIZE);
+        let (connection_update_tx, connection_update_rx) = mpsc::channel(MPSC_CHANNEL_SIZE);
+        let (gossipsub_message_tx, gossipsub_message_rx) = mpsc::channel(MPSC_CHANNEL_SIZE);
+
        // get identity
        let identity = identity.borrow().0.clone();

-        let (to_swarm, from_client) = mpsc::channel(MPSC_CHANNEL_SIZE);
-        let (to_client, from_swarm) = mpsc::channel(MPSC_CHANNEL_SIZE);
        // create networking swarm (within tokio context!! or it crashes)
        let swarm = get_runtime()
-            .block_on(async { Swarm::new(identity, from_client, to_client) })
+            .block_on(async { create_swarm(identity) })
            .pyerr()?;

-        Ok(Self {
-            swarm: Arc::new(Mutex::new(swarm)),
-            from_swarm: Mutex::new(from_swarm),
-            to_swarm: Mutex::new(to_swarm),
-        })
+        // spawn tokio task running the networking logic
+        get_runtime().spawn(async move {
+            networking_task(
+                swarm,
+                to_task_rx,
+                connection_update_tx,
+                gossipsub_message_tx,
+            )
+            .await;
+        });
+        Ok(Self::new(
+            to_task_tx,
+            connection_update_rx,
+            gossipsub_message_rx,
+        ))
    }

    #[gen_stub(skip)]
-    async fn run(&self, #[pyo3(cancel_handle)] mut cancel: CancelHandle) -> PyResult<()> {
-        let copy = Arc::clone(&self.swarm);
-        let jh = get_runtime().spawn(async move {
-            copy.try_lock()
-                .expect("tried to run swarm twice")
-                .run()
-                .await
-        });
-        jh.or(async {
-            cancel.cancelled().await;
-            Ok(())
-        })
-        .await
-        .map_err(|e| PyRuntimeError::new_err(e.to_string()))
+    const fn __traverse__(&self, _visit: PyVisit<'_>) -> Result<(), PyTraverseError> {
+        Ok(()) // This is needed purely so `__clear__` can work
+    }
+
+    #[gen_stub(skip)]
+    fn __clear__(&mut self) {
+        // TODO: may or may not need to await a "kill-signal" oneshot channel message,
+        //       to ensure that the networking task is done BEFORE exiting the clear function...
+        //       but this may require GIL?? and it may not be safe to call GIL here??
+        self.to_task_tx = None; // Using Option<T> as a trick to force channel to be dropped
    }

    // ---- Connection update receiver methods ----

-    /// Receives the next message from networking.
-    async fn recv(&self) -> PyResult<PyMessage> {
-        let msg = pin!(
-            self.from_swarm
-                .try_lock()
-                .expect("called recv concurrently")
-                .recv()
-        )
-        .allow_threads_py()
-        .await;
-        match msg {
-            None => Err(PyConnectionError::new_err("swarm closed")),
-            Some(msg) => msg.try_into(),
-        }
+    /// Receives the next `ConnectionUpdate` from networking.
+    async fn connection_update_recv(&self) -> PyResult<PyConnectionUpdate> {
+        self.connection_update_rx
+            .lock()
+            .allow_threads_py() // allow-threads-aware async call
+            .await
+            .recv_py()
+            .allow_threads_py() // allow-threads-aware async call
+            .await
    }

+    /// Receives at most `limit` `ConnectionUpdate`s from networking and returns them.
+    ///
+    /// For `limit = 0`, an empty collection of `ConnectionUpdate`s will be returned immediately.
+    /// For `limit > 0`, if there are no `ConnectionUpdate`s in the channel's queue this method
+    /// will sleep until a `ConnectionUpdate`s is sent.
+    async fn connection_update_recv_many(&self, limit: usize) -> PyResult<Vec<PyConnectionUpdate>> {
+        self.connection_update_rx
+            .lock()
+            .allow_threads_py() // allow-threads-aware async call
+            .await
+            .recv_many_py(limit)
+            .allow_threads_py() // allow-threads-aware async call
+            .await
+    }
+
+    // TODO: rn this blocks main thread if anything else is awaiting the channel (bc its a mutex)
+    //       so its too dangerous to expose just yet. figure out a better semantics for handling this,
+    //       so things don't randomly block
+    // /// Tries to receive the next `ConnectionUpdate` from networking.
+    // fn connection_update_try_recv(&self) -> PyResult<Option<PyConnectionUpdate>> {
+    //     self.connection_update_rx.blocking_lock().try_recv_py()
+    // }
+    //
+    // /// Checks if the `ConnectionUpdate` channel is empty.
+    // fn connection_update_is_empty(&self) -> bool {
+    //     self.connection_update_rx.blocking_lock().is_empty()
+    // }
+    //
+    // /// Returns the number of `ConnectionUpdate`s in the channel.
+    // fn connection_update_len(&self) -> usize {
+    //     self.connection_update_rx.blocking_lock().len()
+    // }
+
+    // ---- Gossipsub management methods ----
+
    /// Subscribe to a `GossipSub` topic.
-    async fn gossipsub_subscribe(&self, topic: String) -> PyResult<()> {
+    ///
+    /// Returns `True` if the subscription worked. Returns `False` if we were already subscribed.
+    async fn gossipsub_subscribe(&self, topic: String) -> PyResult<bool> {
+        let (tx, rx) = oneshot::channel();
+
        // send off request to subscribe
-        pin!(
-            self.to_swarm
-                .try_lock()
-                .expect("called send concurrently")
-                .send(ToSwarm::Subscribe(topic))
-        )
-        .allow_threads_py() // allow-threads-aware async call
-        .await
-        .map_err(|_| PyConnectionError::new_err("swarm closed"))
+        self.to_task_tx()
+            .send_py(ToTask::GossipsubSubscribe {
+                topic,
+                result_tx: tx,
+            })
+            .allow_threads_py() // allow-threads-aware async call
+            .await?;
+
+        // wait for response & return any errors
+        rx.allow_threads_py() // allow-threads-aware async call
+            .await
+            .map_err(|_| PyErr::receiver_channel_closed())?
    }

    /// Unsubscribes from a `GossipSub` topic.
    ///
    /// Returns `True` if we were subscribed to this topic. Returns `False` if we were not subscribed.
-    async fn gossipsub_unsubscribe(&self, topic: String) -> PyResult<()> {
+    async fn gossipsub_unsubscribe(&self, topic: String) -> PyResult<bool> {
+        let (tx, rx) = oneshot::channel();
+
        // send off request to unsubscribe
-        pin!(
-            self.to_swarm
-                .try_lock()
-                .expect("called send concurrently")
-                .send(ToSwarm::Unsubscribe(topic))
-        )
-        .allow_threads_py() // allow-threads-aware async call
-        .await
-        .map_err(|_| PyConnectionError::new_err("swarm closed"))
+        self.to_task_tx()
+            .send_py(ToTask::GossipsubUnsubscribe {
+                topic,
+                result_tx: tx,
+            })
+            .allow_threads_py() // allow-threads-aware async call
+            .await?;
+
+        // wait for response & convert any errors
+        rx.allow_threads_py() // allow-threads-aware async call
+            .await
+            .map_err(|_| PyErr::receiver_channel_closed())
    }

-    /// Publishes a message to the network on a specific topic.
+    /// Publishes a message with multiple topics to the `GossipSub` network.
+    ///
+    /// If no peers are found that subscribe to this topic, throws `NoPeersSubscribedToTopicError` exception.
    async fn gossipsub_publish(&self, topic: String, data: Py<PyBytes>) -> PyResult<()> {
+        let (tx, rx) = oneshot::channel();
+
        // send off request to subscribe
-        let data = Python::attach(|py| Vec::from(data.as_bytes(py)));
-        pin!(
-            self.to_swarm
-                .try_lock()
-                .expect("called send concurrently")
-                .send(ToSwarm::Message(topic, data))
-        )
-        .allow_threads_py() // allow-threads-aware async call
-        .await
-        .map_err(|_| PyConnectionError::new_err("swarm closed"))
+        let data = Python::with_gil(|py| Vec::from(data.as_bytes(py)));
+        self.to_task_tx()
+            .send_py(ToTask::GossipsubPublish {
+                topic,
+                data,
+                result_tx: tx,
+            })
+            .allow_threads_py() // allow-threads-aware async call
+            .await?;
+
+        // wait for response & return any errors => ignore messageID for now!!!
+        let _ = rx
+            .allow_threads_py() // allow-threads-aware async call
+            .await
+            .map_err(|_| PyErr::receiver_channel_closed())??;
+        Ok(())
    }
+
+    // ---- Gossipsub message receiver methods ----
+
+    /// Receives the next message from the `GossipSub` network.
+    async fn gossipsub_recv(&self) -> PyResult<(String, Py<PyBytes>)> {
+        self.gossipsub_message_rx
+            .lock()
+            .allow_threads_py() // allow-threads-aware async call
+            .await
+            .recv_py()
+            .allow_threads_py() // allow-threads-aware async call
+            .await
+            .map(|(t, d)| (t, d.pybytes()))
+    }
+
+    /// Receives at most `limit` messages from the `GossipSub` network and returns them.
+    ///
+    /// For `limit = 0`, an empty collection of messages will be returned immediately.
+    /// For `limit > 0`, if there are no messages in the channel's queue this method
+    /// will sleep until a message is sent.
+    async fn gossipsub_recv_many(&self, limit: usize) -> PyResult<Vec<(String, Py<PyBytes>)>> {
+        Ok(self
+            .gossipsub_message_rx
+            .lock()
+            .allow_threads_py() // allow-threads-aware async call
+            .await
+            .recv_many_py(limit)
+            .allow_threads_py() // allow-threads-aware async call
+            .await?
+            .into_iter()
+            .map(|(t, d)| (t, d.pybytes()))
+            .collect())
+    }
+
+    // TODO: rn this blocks main thread if anything else is awaiting the channel (bc its a mutex)
+    //       so its too dangerous to expose just yet. figure out a better semantics for handling this,
+    //       so things don't randomly block
+    // /// Tries to receive the next message from the `GossipSub` network.
+    // fn gossipsub_try_recv(&self) -> PyResult<Option<(String, Py<PyBytes>)>> {
+    //     Ok(self
+    //         .gossipsub_message_rx
+    //         .blocking_lock()
+    //         .try_recv_py()?
+    //         .map(|(t, d)| (t, d.pybytes())))
+    // }
+    //
+    // /// Checks if the `GossipSub` message channel is empty.
+    // fn gossipsub_is_empty(&self) -> bool {
+    //     self.gossipsub_message_rx.blocking_lock().is_empty()
+    // }
+    //
+    // /// Returns the number of `GossipSub` messages in the channel.
+    // fn gossipsub_len(&self) -> usize {
+    //     self.gossipsub_message_rx.blocking_lock().len()
+    // }
 }

 pub fn networking_submodule(m: &Bound<'_, PyModule>) -> PyResult<()> {
    m.add_class::<exception::PyNoPeersSubscribedToTopicError>()?;
    m.add_class::<exception::PyAllQueuesFullError>()?;

-    m.add_class::<PySwarm>()?;
-    m.add_class::<PyMessage>()?;
+    m.add_class::<PyConnectionUpdateType>()?;
+    m.add_class::<PyConnectionUpdate>()?;
+    m.add_class::<PyConnectionUpdateType>()?;
+    m.add_class::<PyNetworkingHandle>()?;

    Ok(())
 }
--- a/rust/exo_pyo3_bindings/src/pylibp2p/ident.rs
+++ b/rust/exo_pyo3_bindings/src/pylibp2p/ident.rs
@@ -0,0 +1,159 @@
+use crate::ext::ResultExt as _;
+use libp2p::PeerId;
+use libp2p::identity::Keypair;
+use pyo3::prelude::{PyBytesMethods as _, PyModule, PyModuleMethods as _};
+use pyo3::types::PyBytes;
+use pyo3::{Bound, PyResult, Python, pyclass, pymethods};
+use pyo3_stub_gen::derive::{gen_stub_pyclass, gen_stub_pymethods};
+
+/// Identity keypair of a node.
+#[gen_stub_pyclass]
+#[pyclass(name = "Keypair", frozen)]
+#[repr(transparent)]
+pub struct PyKeypair(pub Keypair);
+
+#[gen_stub_pymethods]
+#[pymethods]
+#[allow(clippy::needless_pass_by_value)]
+impl PyKeypair {
+    /// Generate a new Ed25519 keypair.
+    #[staticmethod]
+    fn generate_ed25519() -> Self {
+        Self(Keypair::generate_ed25519())
+    }
+
+    /// Generate a new ECDSA keypair.
+    #[staticmethod]
+    fn generate_ecdsa() -> Self {
+        Self(Keypair::generate_ecdsa())
+    }
+
+    /// Generate a new Secp256k1 keypair.
+    #[staticmethod]
+    fn generate_secp256k1() -> Self {
+        Self(Keypair::generate_secp256k1())
+    }
+
+    /// Decode a private key from a protobuf structure and parse it as a `Keypair`.
+    #[staticmethod]
+    fn from_protobuf_encoding(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
+        let bytes = Vec::from(bytes.as_bytes());
+        Ok(Self(Keypair::from_protobuf_encoding(&bytes).pyerr()?))
+    }
+
+    /// Decode an keypair from a DER-encoded secret key in PKCS#8 `PrivateKeyInfo`
+    /// format (i.e. unencrypted) as defined in [RFC5208].
+    ///
+    /// [RFC5208]: https://tools.ietf.org/html/rfc5208#section-5
+    #[staticmethod]
+    fn rsa_from_pkcs8(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
+        let mut bytes = Vec::from(bytes.as_bytes());
+        Ok(Self(Keypair::rsa_from_pkcs8(&mut bytes).pyerr()?))
+    }
+
+    /// Decode a keypair from a DER-encoded Secp256k1 secret key in an `ECPrivateKey`
+    /// structure as defined in [RFC5915].
+    ///
+    /// [RFC5915]: https://tools.ietf.org/html/rfc5915
+    #[staticmethod]
+    fn secp256k1_from_der(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
+        let mut bytes = Vec::from(bytes.as_bytes());
+        Ok(Self(Keypair::secp256k1_from_der(&mut bytes).pyerr()?))
+    }
+
+    #[staticmethod]
+    fn ed25519_from_bytes(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
+        let mut bytes = Vec::from(bytes.as_bytes());
+        Ok(Self(Keypair::ed25519_from_bytes(&mut bytes).pyerr()?))
+    }
+
+    /// Encode a private key as protobuf structure.
+    fn to_protobuf_encoding<'py>(&self, py: Python<'py>) -> PyResult<Bound<'py, PyBytes>> {
+        let bytes = self.0.to_protobuf_encoding().pyerr()?;
+        Ok(PyBytes::new(py, &bytes))
+    }
+
+    /// Convert the `Keypair` into the corresponding `PeerId`.
+    fn to_peer_id(&self) -> PyPeerId {
+        PyPeerId(self.0.public().to_peer_id())
+    }
+
+    // /// Hidden constructor for pickling support. TODO: figure out how to do pickling...
+    // #[gen_stub(skip)]
+    // #[new]
+    // fn py_new(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
+    //     Self::from_protobuf_encoding(bytes)
+    // }
+    //
+    // #[gen_stub(skip)]
+    // fn __setstate__(&mut self, state: Bound<'_, PyBytes>) -> PyResult<()> {
+    //     *self = Self::from_protobuf_encoding(state)?;
+    //     Ok(())
+    // }
+    //
+    // #[gen_stub(skip)]
+    // fn __getstate__<'py>(&self, py: Python<'py>) -> PyResult<Bound<'py, PyBytes>> {
+    //     self.to_protobuf_encoding(py)
+    // }
+    //
+    // #[gen_stub(skip)]
+    // pub fn __getnewargs__<'py>(&self, py: Python<'py>) -> PyResult<(Bound<'py, PyBytes>,)> {
+    //     Ok((self.to_protobuf_encoding(py)?,))
+    // }
+}
+
+/// Identifier of a peer of the network.
+///
+/// The data is a `CIDv0` compatible multihash of the protobuf encoded public key of the peer
+/// as specified in [specs/peer-ids](https://github.com/libp2p/specs/blob/master/peer-ids/peer-ids.md).
+#[gen_stub_pyclass]
+#[pyclass(name = "PeerId", frozen)]
+#[derive(Debug, Clone)]
+#[repr(transparent)]
+pub struct PyPeerId(pub PeerId);
+
+#[gen_stub_pymethods]
+#[pymethods]
+#[allow(clippy::needless_pass_by_value)]
+impl PyPeerId {
+    /// Generates a random peer ID from a cryptographically secure PRNG.
+    ///
+    /// This is useful for randomly walking on a DHT, or for testing purposes.
+    #[staticmethod]
+    fn random() -> Self {
+        Self(PeerId::random())
+    }
+
+    /// Parses a `PeerId` from bytes.
+    #[staticmethod]
+    fn from_bytes(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
+        let bytes = Vec::from(bytes.as_bytes());
+        Ok(Self(PeerId::from_bytes(&bytes).pyerr()?))
+    }
+
+    /// Returns a raw bytes representation of this `PeerId`.
+    fn to_bytes<'py>(&self, py: Python<'py>) -> Bound<'py, PyBytes> {
+        let bytes = self.0.to_bytes();
+        PyBytes::new(py, &bytes)
+    }
+
+    /// Returns a base-58 encoded string of this `PeerId`.
+    fn to_base58(&self) -> String {
+        self.0.to_base58()
+    }
+
+    fn __repr__(&self) -> String {
+        format!("PeerId({})", self.to_base58())
+    }
+
+    fn __str__(&self) -> String {
+        self.to_base58()
+    }
+}
+
+pub fn ident_submodule(m: &Bound<'_, PyModule>) -> PyResult<()> {
+    m.add_class::<PyKeypair>()?;
+    m.add_class::<PyPeerId>()?;
+
+    Ok(())
+}
--- a/rust/exo_pyo3_bindings/src/pylibp2p/mod.rs
+++ b/rust/exo_pyo3_bindings/src/pylibp2p/mod.rs
@@ -0,0 +1,8 @@
+//! A module for exposing Rust's libp2p datatypes over Pyo3
+//!
+//! TODO: right now we are coupled to libp2p's identity, but eventually we want to create our own
+//!       independent identity type of some kind or another. This may require handshaking.
+//!
+
+pub mod ident;
+pub mod multiaddr;
--- a/rust/exo_pyo3_bindings/src/pylibp2p/multiaddr.rs
+++ b/rust/exo_pyo3_bindings/src/pylibp2p/multiaddr.rs
@@ -0,0 +1,81 @@
+use crate::ext::ResultExt as _;
+use libp2p::Multiaddr;
+use pyo3::prelude::{PyBytesMethods as _, PyModule, PyModuleMethods as _};
+use pyo3::types::PyBytes;
+use pyo3::{Bound, PyResult, Python, pyclass, pymethods};
+use pyo3_stub_gen::derive::{gen_stub_pyclass, gen_stub_pymethods};
+use std::str::FromStr as _;
+
+/// Representation of a Multiaddr.
+#[gen_stub_pyclass]
+#[pyclass(name = "Multiaddr", frozen)]
+#[derive(Debug, Clone)]
+#[repr(transparent)]
+pub struct PyMultiaddr(pub Multiaddr);
+
+#[gen_stub_pymethods]
+#[pymethods]
+#[allow(clippy::needless_pass_by_value)]
+impl PyMultiaddr {
+    /// Create a new, empty multiaddress.
+    #[staticmethod]
+    fn empty() -> Self {
+        Self(Multiaddr::empty())
+    }
+
+    /// Create a new, empty multiaddress with the given capacity.
+    #[staticmethod]
+    fn with_capacity(n: usize) -> Self {
+        Self(Multiaddr::with_capacity(n))
+    }
+
+    /// Parse a `Multiaddr` value from its byte slice representation.
+    #[staticmethod]
+    fn from_bytes(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
+        let bytes = Vec::from(bytes.as_bytes());
+        Ok(Self(Multiaddr::try_from(bytes).pyerr()?))
+    }
+
+    /// Parse a `Multiaddr` value from its string representation.
+    #[staticmethod]
+    fn from_string(string: String) -> PyResult<Self> {
+        Ok(Self(Multiaddr::from_str(&string).pyerr()?))
+    }
+
+    /// Return the length in bytes of this multiaddress.
+    fn len(&self) -> usize {
+        self.0.len()
+    }
+
+    /// Returns true if the length of this multiaddress is 0.
+    fn is_empty(&self) -> bool {
+        self.0.is_empty()
+    }
+
+    /// Return a copy of this [`Multiaddr`]'s byte representation.
+    fn to_bytes<'py>(&self, py: Python<'py>) -> Bound<'py, PyBytes> {
+        let bytes = self.0.to_vec();
+        PyBytes::new(py, &bytes)
+    }
+
+    /// Convert a Multiaddr to a string.
+    fn to_string(&self) -> String {
+        self.0.to_string()
+    }
+
+    #[gen_stub(skip)]
+    fn __repr__(&self) -> String {
+        format!("Multiaddr({})", self.0)
+    }
+
+    #[gen_stub(skip)]
+    fn __str__(&self) -> String {
+        self.to_string()
+    }
+}
+
+pub fn multiaddr_submodule(m: &Bound<'_, PyModule>) -> PyResult<()> {
+    m.add_class::<PyMultiaddr>()?;
+
+    Ok(())
+}
--- a/rust/networking/Cargo.toml
+++ b/rust/networking/Cargo.toml
@@ -19,14 +19,21 @@ either = { workspace = true }
 # macro dependencies
 extend = { workspace = true }
 delegate = { workspace = true }
+impl-trait-for-tuples = { workspace = true }
+derive_more = { workspace = true }

 # async
 tokio = { workspace = true, features = ["full"] }
-futures-lite = { workspace = true }
+futures = { workspace = true }
 futures-timer = { workspace = true }

 # utility dependencies
 util = { workspace = true }
+thiserror = { workspace = true }
+#internment = { workspace = true }
+#recursion = { workspace = true }
+#generativity = { workspace = true }
+#itertools = { workspace = true }
 tracing-subscriber = { version = "0.3.19", features = ["default", "env-filter"] }
 keccak-const = { workspace = true }

@@ -34,4 +41,4 @@ keccak-const = { workspace = true }
 log = { workspace = true }

 # networking
-libp2p = { workspace = true, features = ["full"] }
+libp2p = { workspace = true, features = ["full"] }
--- a/rust/networking/examples/chatroom.rs
+++ b/rust/networking/examples/chatroom.rs
@@ -1,6 +1,6 @@
-use libp2p::identity;
-use networking::swarm::{FromSwarm, Swarm, ToSwarm};
-use tokio::sync::mpsc;
+use futures::stream::StreamExt as _;
+use libp2p::{gossipsub, identity, swarm::SwarmEvent};
+use networking::{discovery, swarm};
 use tokio::{io, io::AsyncBufReadExt as _, select};
 use tracing_subscriber::EnvFilter;
 use tracing_subscriber::filter::LevelFilter;
@@ -11,50 +11,60 @@ async fn main() {
        .with_env_filter(EnvFilter::from_default_env().add_directive(LevelFilter::INFO.into()))
        .try_init();

-    let (to_swarm, from_client) = mpsc::channel(20);
-    let (to_client, mut from_swarm) = mpsc::channel(20);
    // Configure swarm
-    let mut swarm = Swarm::new(
-        identity::Keypair::generate_ed25519(),
-        from_client,
-        to_client,
-    )
-    .expect("Swarm creation failed");
+    let mut swarm =
+        swarm::create_swarm(identity::Keypair::generate_ed25519()).expect("Swarm creation failed");

    // Create a Gossipsub topic & subscribe
-    _ = to_swarm
-        .send(ToSwarm::Subscribe("test-net".to_owned()))
-        .await;
+    let topic = gossipsub::IdentTopic::new("test-net");
+    swarm
+        .behaviour_mut()
+        .gossipsub
+        .subscribe(&topic)
+        .expect("Subscribing to topic failed");

    // Read full lines from stdin
    let mut stdin = io::BufReader::new(io::stdin()).lines();
    println!("Enter messages via STDIN and they will be sent to connected peers using Gossipsub");

-    tokio::task::spawn(async move { swarm.run().await });
-
    // Kick it off
    loop {
        select! {
            // on gossipsub outgoing
            Ok(Some(line)) = stdin.next_line() => {
-                _= to_swarm.send(ToSwarm::Message("test-net".to_owned(), line.into_bytes())).await;
-            }
-            event = from_swarm.recv() => match event {
-                // on gossipsub incoming
-                Some(FromSwarm::Message(pid, topic, content)) => {
-                    assert_eq!(topic, "test-net");
-                    let fmt = String::from_utf8_lossy(&content);
-                    println!("{pid}: {fmt}");
+                if let Err(e) = swarm
+                    .behaviour_mut().gossipsub
+                    .publish(topic.clone(), line.as_bytes()) {
+                    println!("Publish error: {e:?}");
                }
+            }
+            event = swarm.select_next_some() => match event {
+                // on gossipsub incoming
+                SwarmEvent::Behaviour(swarm::BehaviourEvent::Gossipsub(gossipsub::Event::Message {
+                    propagation_source: peer_id,
+                    message_id: id,
+                    message,
+                })) => println!(
+                        "\n\nGot message: '{}' with id: {id} from peer: {peer_id}\n\n",
+                        String::from_utf8_lossy(&message.data),
+                    ),

                // on discovery
-                Some(FromSwarm::Discovered(pid)) => {
-                        eprintln!("\n\nConnected to: {pid}\n\n");
+                SwarmEvent::Behaviour(swarm::BehaviourEvent::Discovery(e)) => match e {
+                    discovery::Event::ConnectionEstablished {
+                        peer_id, connection_id, remote_ip, remote_tcp_port
+                    } => {
+                        println!("\n\nConnected to: {peer_id}; connection ID: {connection_id}; remote IP: {remote_ip}; remote TCP port: {remote_tcp_port}\n\n");
+                    }
+                    discovery::Event::ConnectionClosed {
+                        peer_id, connection_id, remote_ip, remote_tcp_port
+                    } => {
+                        eprintln!("\n\nDisconnected from: {peer_id}; connection ID: {connection_id}; remote IP: {remote_ip}; remote TCP port: {remote_tcp_port}\n\n");
                    }
-                Some(FromSwarm::Expired(pid)) => {
-                        eprintln!("\n\nDisconnected from: {pid}\n\n");
                }
-                None => break,
+
+                // ignore outgoing errors: those are normal
+                e@SwarmEvent::OutgoingConnectionError { .. } => { log::debug!("Outgoing connection error: {e:?}"); }

                // otherwise log any other event
                e => { log::info!("Other event {e:?}"); }
--- a/rust/networking/examples/chatroom_manual.rs
+++ b/rust/networking/examples/chatroom_manual.rs
@@ -0,0 +1,127 @@
+// Copyright 2018 Parity Technologies (UK) Ltd.
+//
+// Permission is hereby granted, free of charge, to any person obtaining a
+// copy of this software and associated documentation files (the "Software"),
+// to deal in the Software without restriction, including without limitation
+// the rights to use, copy, modify, merge, publish, distribute, sublicense,
+// and/or sell copies of the Software, and to permit persons to whom the
+// Software is furnished to do so, subject to the following conditions:
+//
+// The above copyright notice and this permission notice shall be included in
+// all copies or substantial portions of the Software.
+//
+// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+// OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+// FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+// DEALINGS IN THE SOFTWARE.
+
+use futures::stream::StreamExt;
+use libp2p::{
+    gossipsub, mdns, noise,
+    swarm::{NetworkBehaviour, SwarmEvent},
+    tcp, yamux,
+};
+use std::time::Duration;
+use std::{error::Error, hash::Hash};
+use tokio::{io, io::AsyncBufReadExt, select};
+use tracing_subscriber::EnvFilter;
+
+// We create a custom network behaviour that combines Gossipsub and Mdns.
+#[derive(NetworkBehaviour)]
+struct MyBehaviour {
+    gossipsub: gossipsub::Behaviour,
+    mdns: mdns::tokio::Behaviour,
+}
+
+#[tokio::main]
+async fn main() -> Result<(), Box<dyn Error>> {
+    let _ = tracing_subscriber::fmt()
+        .with_env_filter(EnvFilter::from_default_env())
+        .try_init();
+
+    let mut swarm = libp2p::SwarmBuilder::with_new_identity()
+        .with_tokio()
+        .with_tcp(
+            tcp::Config::default(),
+            noise::Config::new,
+            yamux::Config::default,
+        )?
+        .with_behaviour(|key| {
+            // Set a custom gossipsub configuration
+            let gossipsub_config = gossipsub::ConfigBuilder::default()
+                .heartbeat_interval(Duration::from_secs(10))
+                .validation_mode(gossipsub::ValidationMode::Strict) // This sets the kind of message validation. The default is Strict (enforce message signing)
+                .build()
+                .map_err(io::Error::other)?; // Temporary hack because `build` does not return a proper `std::error::Error`.
+
+            // build a gossipsub network behaviour
+            let gossipsub = gossipsub::Behaviour::new(
+                gossipsub::MessageAuthenticity::Signed(key.clone()),
+                gossipsub_config,
+            )?;
+
+            let mdns =
+                mdns::tokio::Behaviour::new(mdns::Config::default(), key.public().to_peer_id())?;
+            Ok(MyBehaviour { gossipsub, mdns })
+        })?
+        .build();
+
+    println!("Running swarm with identity {}", swarm.local_peer_id());
+
+    // Create a Gossipsub topic
+    let topic = gossipsub::IdentTopic::new("test-net");
+    // subscribes to our topic
+    swarm.behaviour_mut().gossipsub.subscribe(&topic)?;
+
+    // Read full lines from stdin
+    let mut stdin = io::BufReader::new(io::stdin()).lines();
+
+    // Listen on all interfaces and whatever port the OS assigns
+    swarm.listen_on("/ip4/0.0.0.0/tcp/0".parse()?)?;
+
+    println!("Enter messages via STDIN and they will be sent to connected peers using Gossipsub");
+
+    // Kick it off
+    loop {
+        select! {
+            Ok(Some(line)) = stdin.next_line() => {
+                if let Err(e) = swarm
+                    .behaviour_mut().gossipsub
+                    .publish(topic.clone(), line.as_bytes()) {
+                    println!("Publish error: {e:?}");
+                }
+            }
+            event = swarm.select_next_some() => match event {
+                SwarmEvent::Behaviour(MyBehaviourEvent::Mdns(mdns::Event::Discovered(list))) => {
+                    for (peer_id, multiaddr) in list {
+                        println!("mDNS discovered a new peer: {peer_id} on {multiaddr}");
+                        swarm.behaviour_mut().gossipsub.add_explicit_peer(&peer_id);
+                    }
+                },
+                SwarmEvent::Behaviour(MyBehaviourEvent::Mdns(mdns::Event::Expired(list))) => {
+                    for (peer_id, multiaddr) in list {
+                        println!("mDNS discover peer has expired: {peer_id} on {multiaddr}");
+                        swarm.behaviour_mut().gossipsub.remove_explicit_peer(&peer_id);
+                    }
+                },
+                SwarmEvent::Behaviour(MyBehaviourEvent::Gossipsub(gossipsub::Event::Message {
+                    propagation_source: peer_id,
+                    message_id: id,
+                    message,
+                })) => println!(
+                        "Got message: '{}' with id: {id} from peer: {peer_id}",
+                        String::from_utf8_lossy(&message.data),
+                    ),
+                SwarmEvent::NewListenAddr { address, .. } => {
+                    println!("Local node is listening on {address}");
+                }
+                e => {
+                    println!("Other swarm event: {:?}", e);
+                }
+            }
+        }
+    }
+}
--- a/rust/networking/src/discovery.rs
+++ b/rust/networking/src/discovery.rs
@@ -1,10 +1,11 @@
 use crate::ext::MultiaddrExt;
+use crate::keep_alive;
 use delegate::delegate;
 use either::Either;
+use futures::FutureExt;
 use futures_timer::Delay;
 use libp2p::core::transport::PortUse;
 use libp2p::core::{ConnectedPoint, Endpoint};
-use libp2p::futures::FutureExt;
 use libp2p::swarm::behaviour::ConnectionEstablished;
 use libp2p::swarm::dial_opts::DialOpts;
 use libp2p::swarm::{
--- a/rust/networking/src/keep_alive.rs
+++ b/rust/networking/src/keep_alive.rs
@@ -0,0 +1,44 @@
+use delegate::delegate;
+use libp2p::swarm::handler::ConnectionEvent;
+use libp2p::swarm::{ConnectionHandlerEvent, SubstreamProtocol, dummy, handler};
+use std::task::{Context, Poll};
+
+/// An implementation of [`ConnectionHandler`] that doesn't handle any protocols, but it keeps
+/// the connection alive.
+#[derive(Clone)]
+#[repr(transparent)]
+pub struct ConnectionHandler(dummy::ConnectionHandler);
+
+impl ConnectionHandler {
+    pub fn new() -> Self {
+        ConnectionHandler(dummy::ConnectionHandler)
+    }
+}
+
+impl handler::ConnectionHandler for ConnectionHandler {
+    // delegate types and implementation mostly to dummy handler
+    type FromBehaviour = <dummy::ConnectionHandler as handler::ConnectionHandler>::FromBehaviour;
+    type ToBehaviour = <dummy::ConnectionHandler as handler::ConnectionHandler>::ToBehaviour;
+    type InboundProtocol =
+        <dummy::ConnectionHandler as handler::ConnectionHandler>::InboundProtocol;
+    type OutboundProtocol =
+        <dummy::ConnectionHandler as handler::ConnectionHandler>::OutboundProtocol;
+    type InboundOpenInfo =
+        <dummy::ConnectionHandler as handler::ConnectionHandler>::InboundOpenInfo;
+    type OutboundOpenInfo =
+        <dummy::ConnectionHandler as handler::ConnectionHandler>::OutboundOpenInfo;
+
+    delegate! {
+        to self.0 {
+            fn listen_protocol(&self) -> SubstreamProtocol<Self::InboundProtocol, Self::InboundOpenInfo>;
+            fn poll(&mut self, cx: &mut Context<'_>) -> Poll<ConnectionHandlerEvent<Self::OutboundProtocol, Self::OutboundOpenInfo, Self::ToBehaviour>>;
+            fn on_behaviour_event(&mut self, event: Self::FromBehaviour);
+            fn on_connection_event(&mut self, event: ConnectionEvent<Self::InboundProtocol, Self::OutboundProtocol, Self::InboundOpenInfo, Self::OutboundOpenInfo>);
+        }
+    }
+
+    // specifically override this to force connection to stay alive
+    fn connection_keep_alive(&self) -> bool {
+        true
+    }
+}
--- a/rust/networking/src/lib.rs
+++ b/rust/networking/src/lib.rs
@@ -4,7 +4,18 @@
 //!
 //!

+// enable Rust-unstable features for convenience
+#![feature(trait_alias)]
+// #![feature(stmt_expr_attributes)]
+// #![feature(unboxed_closures)]
+// #![feature(assert_matches)]
+// #![feature(async_fn_in_dyn_trait)]
+// #![feature(async_for_loop)]
+// #![feature(auto_traits)]
+// #![feature(negative_impls)]
+
 pub mod discovery;
+pub mod keep_alive;
 pub mod swarm;

 /// Namespace for all the type/trait aliases used by this crate.
@@ -43,3 +54,11 @@ pub(crate) mod ext {
        }
    }
 }
+
+pub(crate) mod private {
+    #![allow(dead_code)]
+
+    /// Sealed traits support
+    pub trait Sealed {}
+    impl<T: ?Sized> Sealed for T {}
+}
--- a/rust/networking/src/swarm.rs
+++ b/rust/networking/src/swarm.rs
@@ -1,30 +1,9 @@
 use crate::alias;
-use crate::discovery;
 use crate::swarm::transport::tcp_transport;
-use behaviour::{Behaviour, BehaviourEvent};
-use futures_lite::StreamExt;
-use libp2p::{PeerId, SwarmBuilder, gossipsub, identity, swarm::SwarmEvent};
-use tokio::sync::mpsc;
+pub use behaviour::{Behaviour, BehaviourEvent};
+use libp2p::{SwarmBuilder, identity};

-pub struct Swarm {
-    swarm: libp2p::Swarm<Behaviour>,
-    from_client: mpsc::Receiver<ToSwarm>,
-    to_client: mpsc::Sender<FromSwarm>,
-}
-
-#[derive(Debug)]
-pub enum FromSwarm {
-    PublishError(gossipsub::PublishError),
-    Discovered(PeerId),
-    Expired(PeerId),
-    Message(PeerId, String, Vec<u8>),
-}
-#[derive(Debug)]
-pub enum ToSwarm {
-    Message(String, Vec<u8>),
-    Subscribe(String),
-    Unsubscribe(String),
-}
+pub type Swarm = libp2p::Swarm<Behaviour>;

 /// The current version of the network: this prevents devices running different versions of the
 /// software from interacting with each other.
@@ -36,142 +15,23 @@ pub enum ToSwarm {
 pub const NETWORK_VERSION: &[u8] = b"v0.0.1";
 pub const OVERRIDE_VERSION_ENV_VAR: &str = "EXO_LIBP2P_NAMESPACE";

-impl Swarm {
-    /// Create and configure a swarm which listens to all ports on OS
-    pub fn new(
-        keypair: identity::Keypair,
-        from_client: mpsc::Receiver<ToSwarm>,
-        to_client: mpsc::Sender<FromSwarm>,
-    ) -> alias::AnyResult<Swarm> {
-        let mut swarm = SwarmBuilder::with_existing_identity(keypair)
-            .with_tokio()
-            .with_other_transport(tcp_transport)?
-            .with_behaviour(Behaviour::new)?
-            .build();
+/// Create and configure a swarm which listens to all ports on OS
+pub fn create_swarm(keypair: identity::Keypair) -> alias::AnyResult<Swarm> {
+    let mut swarm = SwarmBuilder::with_existing_identity(keypair)
+        .with_tokio()
+        .with_other_transport(tcp_transport)?
+        .with_behaviour(Behaviour::new)?
+        .build();

-        // Listen on all interfaces and whatever port the OS assigns
-        swarm.listen_on("/ip4/0.0.0.0/tcp/0".parse()?)?;
-        Ok(Self {
-            swarm,
-            from_client,
-            to_client,
-        })
-    }
-    pub async fn run(&mut self) {
-        log::info!("RUST: networking task started");
-
-        loop {
-            tokio::select! {
-                message = self.from_client.recv() => {
-                    // handle closed channel
-                    let Some(message) = message else {
-                        log::info!("RUST: channel closed");
-                        break;
-                    };
-
-                    // dispatch incoming messages
-                    match message {
-                        ToSwarm::Subscribe(topic) => {
-                            // try to subscribe
-                            match self.swarm.behaviour_mut().gossipsub.subscribe(&gossipsub::IdentTopic::new(topic.clone())) {
-                                    Err(e) => {
-                                        let gossipsub::SubscriptionError::PublishError(e) = e else {
-                                            unreachable!("topic filter used")
-                                        };
-                                        let Ok(()) = self.to_client.send(FromSwarm::PublishError(e)).await else {
-                                            log::warn!("RUST: client connection closed");
-                                            break
-                                        };
-                                    },
-                                    Ok(false) => log::warn!("RUST: tried to subscribe to topic twice"),
-                                    Ok(true) => {},
-                                }
-                        }
-                        ToSwarm::Unsubscribe(topic) => {
-                            // try to subscribe
-                            if !self.swarm.behaviour_mut().gossipsub.unsubscribe(&gossipsub::IdentTopic::new(topic)) {
-                                log::warn!("RUST: tried to unsubscribe from topic twice");
-                            }
-                        }
-                        ToSwarm::Message( topic, data ) => {
-                            // try to publish the data -> catch NoPeersSubscribedToTopic error & convert to correct exception
-                            match self.swarm.behaviour_mut().gossipsub.publish(
-                                gossipsub::IdentTopic::new(topic), data
-                            ) {
-                                Ok(_) => {},
-                                Err(e) => {
-                                    let Ok(()) = self.to_client.send(FromSwarm::PublishError(e)).await else {
-                                        log::warn!("RUST: client connection closed");
-                                        break
-                                    };
-                                },
-                            }
-                        }
-                    }
-                }
-
-                // architectural solution to this problem:
-                // create keep_alive behavior who's job it is to dial peers discovered by mDNS (and drop when expired)
-                //   -> it will emmit TRUE connected/disconnected events consumable elsewhere
-                //
-                // gossipsub will feed off-of dial attempts created by networking, and that will bootstrap its' peers list
-                // then for actual communication it will dial those peers if need-be
-                swarm_event = self.swarm.next() => {
-                    let Some(swarm_event) = swarm_event else {
-                        log::warn!("RUST: swarm closed communication");
-                        break
-                    };
-                    let SwarmEvent::Behaviour(behaviour_event) = swarm_event else {
-                        continue
-                    };
-                    match behaviour_event {
-                        BehaviourEvent::Gossipsub(gossipsub::Event::Message {
-                            message: gossipsub::Message {
-                                source,
-                                topic,
-                                data,
-                                ..
-                            },
-                            ..
-                        }) => {
-                            let Some(peer_id) = source else {
-                                log::warn!("RUST: ignoring message with unknown source on {topic}");
-                                continue;
-                            };
-                            // send incoming message to channel (or exit if connection closed)
-                            if let Err(e) = self.to_client.send(FromSwarm::Message(peer_id, topic.into_string(), data)).await {
-                                log::warn!("RUST: could not send incoming gossipsub message since channel already closed: {e}");
-                                break
-                            };
-                        },
-                        BehaviourEvent::Discovery(discovery::Event::ConnectionEstablished { peer_id, .. }) => {
-                            // send connection event to channel (or exit if connection closed)
-                            if let Err(_) = self.to_client.send(FromSwarm::Discovered(peer_id)).await {
-                                log::warn!("RUST: swarm closed communication");
-                            };
-                        },
-                        BehaviourEvent::Discovery(discovery::Event::ConnectionClosed { peer_id, .. }) => {
-                            // send connection event to channel (or exit if connection closed)
-                            if let Err(_) = self.to_client.send(FromSwarm::Expired(peer_id)).await {
-                                log::warn!("RUST: swarm closed communication");
-                            };
-                        },
-                        e => {
-                            log::debug!("RUST: other event {e:?}");
-                        }
-                    }
-                }
-            }
-        }
-
-        log::info!("RUST: networking task stopped");
-    }
+    // Listen on all interfaces and whatever port the OS assigns
+    swarm.listen_on("/ip4/0.0.0.0/tcp/0".parse()?)?;
+    Ok(swarm)
 }

 mod transport {
    use crate::alias;
    use crate::swarm::{NETWORK_VERSION, OVERRIDE_VERSION_ENV_VAR};
-    use futures_lite::{AsyncRead, AsyncWrite};
+    use futures::{AsyncRead, AsyncWrite};
    use keccak_const::Sha3_256;
    use libp2p::core::muxing;
    use libp2p::core::transport::Boxed;
--- a/src/exo/download/coordinator.py
+++ b/src/exo/download/coordinator.py
@@ -14,7 +14,6 @@ from exo.download.download_utils import (
    map_repo_download_progress_to_download_progress_data,
 )
 from exo.download.shard_downloader import ShardDownloader
-from exo.shared.constants import EXO_MODELS_DIR
 from exo.shared.models.model_cards import ModelId
 from exo.shared.types.commands import (
    CancelDownload,
@@ -64,9 +63,6 @@ class DownloadCoordinator:
        self.event_sender, self.event_receiver = channel[Event]()
        self.shard_downloader.on_progress(self._download_progress_callback)

-    def _model_dir(self, model_id: ModelId) -> str:
-        return str(EXO_MODELS_DIR / model_id.normalize())
-
    async def _download_progress_callback(
        self, callback_shard: ShardMetadata, progress: RepoDownloadProgress
    ) -> None:
@@ -78,7 +74,6 @@ class DownloadCoordinator:
                shard_metadata=callback_shard,
                node_id=self.node_id,
                total_bytes=progress.total_bytes,
-                model_directory=self._model_dir(model_id),
            )
            self.download_status[model_id] = completed
            await self.event_sender.send(
@@ -98,7 +93,6 @@ class DownloadCoordinator:
                download_progress=map_repo_download_progress_to_download_progress_data(
                    progress
                ),
-                model_directory=self._model_dir(model_id),
            )
            self.download_status[model_id] = ongoing
            await self.event_sender.send(
@@ -176,11 +170,7 @@ class DownloadCoordinator:
                return

        # Emit pending status
-        progress = DownloadPending(
-            shard_metadata=shard,
-            node_id=self.node_id,
-            model_directory=self._model_dir(model_id),
-        )
+        progress = DownloadPending(shard_metadata=shard, node_id=self.node_id)
        self.download_status[model_id] = progress
        await self.event_sender.send(NodeDownloadProgress(download_progress=progress))

@@ -194,7 +184,6 @@ class DownloadCoordinator:
                shard_metadata=shard,
                node_id=self.node_id,
                total_bytes=initial_progress.total_bytes,
-                model_directory=self._model_dir(model_id),
            )
            self.download_status[model_id] = completed
            await self.event_sender.send(
@@ -217,7 +206,6 @@ class DownloadCoordinator:
            download_progress=map_repo_download_progress_to_download_progress_data(
                initial_progress
            ),
-            model_directory=self._model_dir(model_id),
        )
        self.download_status[model_id] = status
        self.event_sender.send_nowait(NodeDownloadProgress(download_progress=status))
@@ -231,7 +219,6 @@ class DownloadCoordinator:
                    shard_metadata=shard,
                    node_id=self.node_id,
                    error_message=str(e),
-                    model_directory=self._model_dir(model_id),
                )
                self.download_status[model_id] = failed
                await self.event_sender.send(
@@ -266,7 +253,6 @@ class DownloadCoordinator:
            pending = DownloadPending(
                shard_metadata=current_status.shard_metadata,
                node_id=self.node_id,
-                model_directory=self._model_dir(model_id),
            )
            await self.event_sender.send(
                NodeDownloadProgress(download_progress=pending)
@@ -309,18 +295,11 @@ class DownloadCoordinator:
                            node_id=self.node_id,
                            shard_metadata=progress.shard,
                            total_bytes=progress.total_bytes,
-                            model_directory=self._model_dir(
-                                progress.shard.model_card.model_id
-                            ),
                        )
                    elif progress.status in ["in_progress", "not_started"]:
                        if progress.downloaded_bytes_this_session.in_bytes == 0:
                            status = DownloadPending(
-                                node_id=self.node_id,
-                                shard_metadata=progress.shard,
-                                model_directory=self._model_dir(
-                                    progress.shard.model_card.model_id
-                                ),
+                                node_id=self.node_id, shard_metadata=progress.shard
                            )
                        else:
                            status = DownloadOngoing(
@@ -329,9 +308,6 @@ class DownloadCoordinator:
                                download_progress=map_repo_download_progress_to_download_progress_data(
                                    progress
                                ),
-                                model_directory=self._model_dir(
-                                    progress.shard.model_card.model_id
-                                ),
                            )
                    else:
                        continue
--- a/src/exo/main.py
+++ b/src/exo/main.py
@@ -44,7 +44,7 @@ class Node:
    @classmethod
    async def create(cls, args: "Args") -> "Self":
        keypair = get_node_id_keypair()
-        node_id = NodeId(keypair.to_string())
+        node_id = NodeId(keypair.to_peer_id().to_base58())
        session_id = SessionId(master_node_id=node_id, election_clock=0)
        router = Router.create(keypair)
        await router.register_topic(topics.GLOBAL_EVENTS)
@@ -136,8 +136,6 @@ class Node:

    async def run(self):
        async with self._tg as tg:
-            signal.signal(signal.SIGINT, lambda _, __: self.shutdown())
-            signal.signal(signal.SIGTERM, lambda _, __: self.shutdown())
            tg.start_soon(self.router.run)
            tg.start_soon(self.election.run)
            if self.download_coordinator:
@@ -149,6 +147,8 @@ class Node:
            if self.api:
                tg.start_soon(self.api.run)
            tg.start_soon(self._elect_loop)
+            signal.signal(signal.SIGINT, lambda _, __: self.shutdown())
+            signal.signal(signal.SIGTERM, lambda _, __: self.shutdown())

    def shutdown(self):
        # if this is our second call to shutdown, just sys.exit
--- a/src/exo/master/adapters/chat_completions.py
+++ b/src/exo/master/adapters/chat_completions.py
@@ -17,9 +17,13 @@ from exo.shared.types.api import (
    LogprobsContentItem,
    StreamingChoiceResponse,
    ToolCall,
-    Usage,
 )
-from exo.shared.types.chunks import ErrorChunk, TokenChunk, ToolCallChunk
+from exo.shared.types.chunks import (
+    ErrorChunk,
+    PrefillProgressData,
+    TokenChunk,
+    ToolCallChunk,
+)
 from exo.shared.types.common import CommandId
 from exo.shared.types.text_generation import InputMessage, TextGenerationTaskParams

@@ -123,70 +127,71 @@ def chunk_to_response(

 async def generate_chat_stream(
    command_id: CommandId,
-    chunk_stream: AsyncGenerator[ErrorChunk | ToolCallChunk | TokenChunk, None],
+    event_stream: AsyncGenerator[
+        PrefillProgressData | ErrorChunk | ToolCallChunk | TokenChunk, None
+    ],
 ) -> AsyncGenerator[str, None]:
-    """Generate Chat Completions API streaming events from chunks."""
-    last_usage: Usage | None = None
+    """Generate Chat Completions API streaming events from StreamEvents.

-    async for chunk in chunk_stream:
-        if isinstance(chunk, ErrorChunk):
-            error_response = ErrorResponse(
-                error=ErrorInfo(
-                    message=chunk.error_message or "Internal server error",
-                    type="InternalServerError",
-                    code=500,
-                )
-            )
-            yield f"data: {error_response.model_dump_json()}\n\n"
-            yield "data: [DONE]\n\n"
-            return
+    Handles PrefillProgressData, ErrorChunk, ToolCallChunk, and TokenChunk.
+    """
+    async for event in event_stream:
+        match event:
+            case PrefillProgressData():
+                yield f"event: prefill_progress\ndata: {event.model_dump_json()}\n\n"

-        last_usage = chunk.usage or last_usage
-
-        if isinstance(chunk, ToolCallChunk):
-            tool_call_deltas = [
-                ToolCall(
-                    id=tool.id,
-                    index=i,
-                    function=tool,
-                )
-                for i, tool in enumerate(chunk.tool_calls)
-            ]
-            tool_response = ChatCompletionResponse(
-                id=command_id,
-                created=int(time.time()),
-                model=chunk.model,
-                choices=[
-                    StreamingChoiceResponse(
-                        index=0,
-                        delta=ChatCompletionMessage(
-                            role="assistant",
-                            tool_calls=tool_call_deltas,
-                        ),
-                        finish_reason="tool_calls",
+            case ErrorChunk():
+                error_response = ErrorResponse(
+                    error=ErrorInfo(
+                        message=event.error_message or "Internal server error",
+                        type="InternalServerError",
+                        code=500,
                    )
-                ],
-                usage=last_usage,
-            )
-            yield f"data: {tool_response.model_dump_json()}\n\n"
-            yield "data: [DONE]\n\n"
-            return
+                )
+                yield f"data: {error_response.model_dump_json()}\n\n"
+                yield "data: [DONE]\n\n"
+                return

-        chunk_response = chunk_to_response(chunk, command_id)
-        if chunk.finish_reason is not None:
-            chunk_response = chunk_response.model_copy(update={"usage": last_usage})
-        yield f"data: {chunk_response.model_dump_json()}\n\n"
+            case ToolCallChunk():
+                tool_call_deltas = [
+                    ToolCall(
+                        id=tool.id,
+                        index=i,
+                        function=tool,
+                    )
+                    for i, tool in enumerate(event.tool_calls)
+                ]
+                tool_response = ChatCompletionResponse(
+                    id=command_id,
+                    created=int(time.time()),
+                    model=event.model,
+                    choices=[
+                        StreamingChoiceResponse(
+                            index=0,
+                            delta=ChatCompletionMessage(
+                                role="assistant",
+                                tool_calls=tool_call_deltas,
+                            ),
+                            finish_reason="tool_calls",
+                        )
+                    ],
+                )
+                yield f"data: {tool_response.model_dump_json()}\n\n"
+                yield "data: [DONE]\n\n"
+                return

-        if chunk.finish_reason is not None:
-            yield "data: [DONE]\n\n"
+            case TokenChunk():
+                chunk_response = chunk_to_response(event, command_id)
+                yield f"data: {chunk_response.model_dump_json()}\n\n"
+
+                if event.finish_reason is not None:
+                    yield "data: [DONE]\n\n"


 async def collect_chat_response(
    command_id: CommandId,
    chunk_stream: AsyncGenerator[ErrorChunk | ToolCallChunk | TokenChunk, None],
-) -> AsyncGenerator[str]:
-    # This is an AsyncGenerator[str] rather than returning a ChatCompletionReponse because
-    # FastAPI handles the cancellation better but wouldn't auto-serialize for some reason
+) -> ChatCompletionResponse:
    """Collect all token chunks and return a single ChatCompletionResponse."""
    text_parts: list[str] = []
    tool_calls: list[ToolCall] = []
@@ -194,7 +199,6 @@ async def collect_chat_response(
    model: str | None = None
    finish_reason: FinishReason | None = None
    error_message: str | None = None
-    last_usage: Usage | None = None

    async for chunk in chunk_stream:
        if isinstance(chunk, ErrorChunk):
@@ -204,8 +208,6 @@ async def collect_chat_response(
        if model is None:
            model = chunk.model

-        last_usage = chunk.usage or last_usage
-
        if isinstance(chunk, TokenChunk):
            text_parts.append(chunk.text)
            if chunk.logprob is not None:
@@ -236,7 +238,7 @@ async def collect_chat_response(
    combined_text = "".join(text_parts)
    assert model is not None

-    yield ChatCompletionResponse(
+    return ChatCompletionResponse(
        id=command_id,
        created=int(time.time()),
        model=model,
@@ -254,6 +256,4 @@ async def collect_chat_response(
                finish_reason=finish_reason,
            )
        ],
-        usage=last_usage,
-    ).model_dump_json()
-    return
+    )
--- a/src/exo/master/adapters/claude.py
+++ b/src/exo/master/adapters/claude.py
@@ -4,7 +4,7 @@ import json
 from collections.abc import AsyncGenerator
 from typing import Any

-from exo.shared.types.api import FinishReason, Usage
+from exo.shared.types.api import FinishReason
 from exo.shared.types.chunks import ErrorChunk, TokenChunk, ToolCallChunk
 from exo.shared.types.claude_api import (
    ClaudeContentBlock,
@@ -161,14 +161,12 @@ async def collect_claude_response(
    command_id: CommandId,
    model: str,
    chunk_stream: AsyncGenerator[ErrorChunk | ToolCallChunk | TokenChunk, None],
-) -> AsyncGenerator[str]:
-    # This is an AsyncGenerator[str] rather than returning a ChatCompletionReponse because
-    # FastAPI handles the cancellation better but wouldn't auto-serialize for some reason
+) -> ClaudeMessagesResponse:
    """Collect all token chunks and return a single ClaudeMessagesResponse."""
    text_parts: list[str] = []
    tool_use_blocks: list[ClaudeToolUseBlock] = []
    stop_reason: ClaudeStopReason | None = None
-    last_usage: Usage | None = None
+    last_stats = None
    error_message: str | None = None

    async for chunk in chunk_stream:
@@ -176,8 +174,6 @@ async def collect_claude_response(
            error_message = chunk.error_message or "Internal server error"
            break

-        last_usage = chunk.usage or last_usage
-
        if isinstance(chunk, ToolCallChunk):
            for tool in chunk.tool_calls:
                tool_use_blocks.append(
@@ -187,10 +183,12 @@ async def collect_claude_response(
                        input=json.loads(tool.arguments),  # pyright: ignore[reportAny]
                    )
                )
+            last_stats = chunk.stats or last_stats
            stop_reason = "tool_use"
            continue

        text_parts.append(chunk.text)
+        last_stats = chunk.stats or last_stats

        if chunk.finish_reason is not None:
            stop_reason = finish_reason_to_claude_stop_reason(chunk.finish_reason)
@@ -210,11 +208,11 @@ async def collect_claude_response(
    if not content:
        content.append(ClaudeTextBlock(text=""))

-    # Use actual usage data if available
-    input_tokens = last_usage.prompt_tokens if last_usage else 0
-    output_tokens = last_usage.completion_tokens if last_usage else 0
+    # Use actual usage data from stats if available
+    input_tokens = last_stats.prompt_tokens if last_stats else 0
+    output_tokens = last_stats.generation_tokens if last_stats else 0

-    yield ClaudeMessagesResponse(
+    return ClaudeMessagesResponse(
        id=f"msg_{command_id}",
        model=model,
        content=content,
@@ -223,8 +221,7 @@ async def collect_claude_response(
            input_tokens=input_tokens,
            output_tokens=output_tokens,
        ),
-    ).model_dump_json()
-    return
+    )


 async def generate_claude_stream(
@@ -252,7 +249,7 @@ async def generate_claude_stream(

    output_tokens = 0
    stop_reason: ClaudeStopReason | None = None
-    last_usage: Usage | None = None
+    last_stats = None
    next_block_index = 1  # text block is 0, tool blocks start at 1

    async for chunk in chunk_stream:
@@ -260,9 +257,8 @@ async def generate_claude_stream(
            # Close text block and bail
            break

-        last_usage = chunk.usage or last_usage
-
        if isinstance(chunk, ToolCallChunk):
+            last_stats = chunk.stats or last_stats
            stop_reason = "tool_use"

            # Emit tool_use content blocks
@@ -294,6 +290,7 @@ async def generate_claude_stream(
            continue

        output_tokens += 1  # Count each chunk as one token
+        last_stats = chunk.stats or last_stats

        # content_block_delta
        delta_event = ClaudeContentBlockDeltaEvent(
@@ -305,9 +302,9 @@ async def generate_claude_stream(
        if chunk.finish_reason is not None:
            stop_reason = finish_reason_to_claude_stop_reason(chunk.finish_reason)

-    # Use actual token count from usage if available
-    if last_usage is not None:
-        output_tokens = last_usage.completion_tokens
+    # Use actual token count from stats if available
+    if last_stats is not None:
+        output_tokens = last_stats.generation_tokens

    # content_block_stop for text block
    block_stop = ClaudeContentBlockStopEvent(index=0)
--- a/src/exo/master/adapters/responses.py
+++ b/src/exo/master/adapters/responses.py
@@ -4,7 +4,6 @@ from collections.abc import AsyncGenerator
 from itertools import count
 from typing import Any

-from exo.shared.types.api import Usage
 from exo.shared.types.chunks import ErrorChunk, TokenChunk, ToolCallChunk
 from exo.shared.types.common import CommandId
 from exo.shared.types.openai_responses import (
@@ -122,15 +121,13 @@ async def collect_responses_response(
    command_id: CommandId,
    model: str,
    chunk_stream: AsyncGenerator[ErrorChunk | ToolCallChunk | TokenChunk, None],
-) -> AsyncGenerator[str]:
-    # This is an AsyncGenerator[str] rather than returning a ChatCompletionReponse because
-    # FastAPI handles the cancellation better but wouldn't auto-serialize for some reason
+) -> ResponsesResponse:
    """Collect all token chunks and return a single ResponsesResponse."""
    response_id = f"resp_{command_id}"
    item_id = f"item_{command_id}"
    accumulated_text = ""
    function_call_items: list[ResponseFunctionCallItem] = []
-    last_usage: Usage | None = None
+    last_stats = None
    error_message: str | None = None

    async for chunk in chunk_stream:
@@ -138,32 +135,32 @@ async def collect_responses_response(
            error_message = chunk.error_message or "Internal server error"
            break

-        last_usage = chunk.usage or last_usage
-
        if isinstance(chunk, ToolCallChunk):
            for tool in chunk.tool_calls:
                function_call_items.append(
                    ResponseFunctionCallItem(
-                        id=tool.id,
-                        call_id=tool.id,
+                        id=f"fc_{tool.id}",
+                        call_id=f"call_{tool.id}",
                        name=tool.name,
                        arguments=tool.arguments,
                    )
                )
+            last_stats = chunk.stats or last_stats
            continue

        accumulated_text += chunk.text
+        last_stats = chunk.stats or last_stats

    if error_message is not None:
        raise ValueError(error_message)

-    # Create usage from usage data if available
+    # Create usage from stats if available
    usage = None
-    if last_usage is not None:
+    if last_stats is not None:
        usage = ResponseUsage(
-            input_tokens=last_usage.prompt_tokens,
-            output_tokens=last_usage.completion_tokens,
-            total_tokens=last_usage.total_tokens,
+            input_tokens=last_stats.prompt_tokens,
+            output_tokens=last_stats.generation_tokens,
+            total_tokens=last_stats.prompt_tokens + last_stats.generation_tokens,
        )

    output: list[ResponseItem] = [
@@ -175,15 +172,14 @@ async def collect_responses_response(
    ]
    output.extend(function_call_items)

-    yield ResponsesResponse(
+    return ResponsesResponse(
        id=response_id,
        model=model,
        status="completed",
        output=output,
        output_text=accumulated_text,
        usage=usage,
-    ).model_dump_json()
-    return
+    )


 async def generate_responses_stream(
@@ -239,16 +235,15 @@ async def generate_responses_stream(

    accumulated_text = ""
    function_call_items: list[ResponseFunctionCallItem] = []
-    last_usage: Usage | None = None
+    last_stats = None
    next_output_index = 1  # message item is at 0

    async for chunk in chunk_stream:
        if isinstance(chunk, ErrorChunk):
            break

-        last_usage = chunk.usage or last_usage
-
        if isinstance(chunk, ToolCallChunk):
+            last_stats = chunk.stats or last_stats
            for tool in chunk.tool_calls:
                fc_id = f"fc_{tool.id}"
                call_id = f"call_{tool.id}"
@@ -307,6 +302,7 @@ async def generate_responses_stream(
            continue

        accumulated_text += chunk.text
+        last_stats = chunk.stats or last_stats

        # response.output_text.delta
        delta_event = ResponseTextDeltaEvent(
@@ -350,13 +346,13 @@ async def generate_responses_stream(
    )
    yield f"event: response.output_item.done\ndata: {item_done.model_dump_json()}\n\n"

-    # Create usage from usage data if available
+    # Create usage from stats if available
    usage = None
-    if last_usage is not None:
+    if last_stats is not None:
        usage = ResponseUsage(
-            input_tokens=last_usage.prompt_tokens,
-            output_tokens=last_usage.completion_tokens,
-            total_tokens=last_usage.total_tokens,
+            input_tokens=last_stats.prompt_tokens,
+            output_tokens=last_stats.generation_tokens,
+            total_tokens=last_stats.prompt_tokens + last_stats.generation_tokens,
        )

    # response.completed
--- a/src/exo/master/api.py
+++ b/src/exo/master/api.py
@@ -105,6 +105,7 @@ from exo.shared.types.chunks import (
    ErrorChunk,
    ImageChunk,
    InputImageChunk,
+    PrefillProgressData,
    TokenChunk,
    ToolCallChunk,
 )
@@ -125,7 +126,6 @@ from exo.shared.types.commands import (
    PlaceInstance,
    SendInputChunk,
    StartDownload,
-    TaskCancelled,
    TaskFinished,
    TextGeneration,
 )
@@ -135,6 +135,7 @@ from exo.shared.types.events import (
    Event,
    ForwarderEvent,
    IndexedEvent,
+    PrefillProgress,
    TracesMerged,
 )
 from exo.shared.types.memory import Memory
@@ -218,7 +219,8 @@ class API:
        )

        self._text_generation_queues: dict[
-            CommandId, Sender[TokenChunk | ErrorChunk | ToolCallChunk]
+            CommandId,
+            Sender[TokenChunk | ErrorChunk | ToolCallChunk | PrefillProgressData],
        ] = {}
        self._image_generation_queues: dict[
            CommandId, Sender[ImageChunk | ErrorChunk]
@@ -522,36 +524,51 @@ class API:
            instance_id=instance_id,
        )

-    async def _token_chunk_stream(
+    async def _stream_events(
        self, command_id: CommandId
-    ) -> AsyncGenerator[ErrorChunk | ToolCallChunk | TokenChunk, None]:
-        """Yield chunks for a given command until completion.
+    ) -> AsyncGenerator[
+        TokenChunk | ErrorChunk | ToolCallChunk | PrefillProgressData, None
+    ]:
+        """Yield stream events for a command.

        This is the internal low-level stream used by all API adapters.
        """
        try:
            self._text_generation_queues[command_id], recv = channel[
-                ErrorChunk | ToolCallChunk | TokenChunk
+                TokenChunk | ErrorChunk | ToolCallChunk | PrefillProgressData
            ]()

-            with recv as token_chunks:
-                async for chunk in token_chunks:
-                    yield chunk
-                    if chunk.finish_reason is not None:
+            with recv as events:
+                async for event in events:
+                    yield event
+                    if (
+                        isinstance(event, TokenChunk)
+                        and event.finish_reason is not None
+                    ):
                        break

        except anyio.get_cancelled_exc_class():
-            command = TaskCancelled(cancelled_command_id=command_id)
-            with anyio.CancelScope(shield=True):
-                await self.command_sender.send(
-                    ForwarderCommand(origin=self.node_id, command=command)
-                )
+            # TODO: TaskCancelled
+            """
+            self.command_sender.send_nowait(
+                ForwarderCommand(origin=self.node_id, command=command)
+            )
+            """
            raise
        finally:
-            await self._send(TaskFinished(finished_command_id=command_id))
+            command = TaskFinished(finished_command_id=command_id)
+            await self._send(command)
            if command_id in self._text_generation_queues:
                del self._text_generation_queues[command_id]

+    async def _chunk_stream(
+        self, command_id: CommandId
+    ) -> AsyncGenerator[ErrorChunk | ToolCallChunk | TokenChunk, None]:
+        """Yield chunks, filtering out prefill progress events."""
+        async for event in self._stream_events(command_id):
+            if not isinstance(event, PrefillProgressData):
+                yield event
+
    async def _collect_text_generation_with_stats(
        self, command_id: CommandId
    ) -> BenchChatCompletionResponse:
@@ -562,7 +579,7 @@ class API:

        stats: GenerationStats | None = None

-        async for chunk in self._token_chunk_stream(command_id):
+        async for chunk in self._chunk_stream(command_id):
            if chunk.finish_reason == "error":
                raise HTTPException(
                    status_code=500,
@@ -634,7 +651,7 @@ class API:
            return StreamingResponse(
                generate_chat_stream(
                    command.command_id,
-                    self._token_chunk_stream(command.command_id),
+                    self._stream_events(command.command_id),
                ),
                media_type="text/event-stream",
                headers={
@@ -643,14 +660,14 @@ class API:
                    "X-Accel-Buffering": "no",
                },
            )
-        else:
-            return StreamingResponse(
-                collect_chat_response(
-                    command.command_id,
-                    self._token_chunk_stream(command.command_id),
-                ),
-                media_type="application/json",
+
+        try:
+            return await collect_chat_response(
+                command.command_id,
+                self._chunk_stream(command.command_id),
            )
+        except ValueError as e:
+            raise HTTPException(status_code=500, detail=str(e)) from e

    async def bench_chat_completions(
        self, payload: BenchChatCompletionRequest
@@ -666,7 +683,8 @@ class API:
        command = TextGeneration(task_params=task_params)
        await self._send(command)

-        return await self._collect_text_generation_with_stats(command.command_id)
+        response = await self._collect_text_generation_with_stats(command.command_id)
+        return response

    async def _resolve_and_validate_text_model(self, model_id: ModelId) -> ModelId:
        """Validate a text model exists and return the resolved model ID.
@@ -884,11 +902,6 @@ class API:
                        del image_metadata[key]

        except anyio.get_cancelled_exc_class():
-            command = TaskCancelled(cancelled_command_id=command_id)
-            with anyio.CancelScope(shield=True):
-                await self.command_sender.send(
-                    ForwarderCommand(origin=self.node_id, command=command)
-                )
            raise
        finally:
            await self._send(TaskFinished(finished_command_id=command_id))
@@ -970,11 +983,6 @@ class API:

            return (images, stats if capture_stats else None)
        except anyio.get_cancelled_exc_class():
-            command = TaskCancelled(cancelled_command_id=command_id)
-            with anyio.CancelScope(shield=True):
-                await self.command_sender.send(
-                    ForwarderCommand(origin=self.node_id, command=command)
-                )
            raise
        finally:
            await self._send(TaskFinished(finished_command_id=command_id))
@@ -1223,7 +1231,7 @@ class API:
                generate_claude_stream(
                    command.command_id,
                    payload.model,
-                    self._token_chunk_stream(command.command_id),
+                    self._chunk_stream(command.command_id),
                ),
                media_type="text/event-stream",
                headers={
@@ -1232,15 +1240,15 @@ class API:
                    "X-Accel-Buffering": "no",
                },
            )
-        else:
-            return StreamingResponse(
-                collect_claude_response(
-                    command.command_id,
-                    payload.model,
-                    self._token_chunk_stream(command.command_id),
-                ),
-                media_type="application/json",
+
+        try:
+            return await collect_claude_response(
+                command.command_id,
+                payload.model,
+                self._chunk_stream(command.command_id),
            )
+        except ValueError as e:
+            raise HTTPException(status_code=500, detail=str(e)) from e

    async def openai_responses(
        self, payload: ResponsesRequest
@@ -1258,7 +1266,7 @@ class API:
                generate_responses_stream(
                    command.command_id,
                    payload.model,
-                    self._token_chunk_stream(command.command_id),
+                    self._chunk_stream(command.command_id),
                ),
                media_type="text/event-stream",
                headers={
@@ -1268,15 +1276,14 @@ class API:
                },
            )

-        else:
-            return StreamingResponse(
-                collect_responses_response(
-                    command.command_id,
-                    payload.model,
-                    self._token_chunk_stream(command.command_id),
-                ),
-                media_type="application/json",
+        try:
+            return await collect_responses_response(
+                command.command_id,
+                payload.model,
+                self._chunk_stream(command.command_id),
            )
+        except ValueError as e:
+            raise HTTPException(status_code=500, detail=str(e)) from e

    def _calculate_total_available_memory(self) -> Memory:
        """Calculate total available memory across all nodes in bytes."""
@@ -1430,6 +1437,20 @@ class API:
                            except BrokenResourceError:
                                self._text_generation_queues.pop(event.command_id, None)

+                    elif isinstance(event, PrefillProgress):
+                        if queue := self._text_generation_queues.get(
+                            event.command_id, None
+                        ):
+                            try:
+                                await queue.send(
+                                    PrefillProgressData(
+                                        processed_tokens=event.processed_tokens,
+                                        total_tokens=event.total_tokens,
+                                    )
+                                )
+                            except BrokenResourceError:
+                                self._text_generation_queues.pop(event.command_id, None)
+
                    if isinstance(event, TracesMerged):
                        self._save_merged_trace(event)

--- a/src/exo/master/main.py
+++ b/src/exo/master/main.py
@@ -24,7 +24,6 @@ from exo.shared.types.commands import (
    PlaceInstance,
    RequestEventLog,
    SendInputChunk,
-    TaskCancelled,
    TaskFinished,
    TestCommand,
    TextGeneration,
@@ -40,7 +39,6 @@ from exo.shared.types.events import (
    NodeTimedOut,
    TaskCreated,
    TaskDeleted,
-    TaskStatusUpdated,
    TraceEventData,
    TracesCollected,
    TracesMerged,
@@ -281,7 +279,7 @@ class Master:
                        case DeleteInstance():
                            placement = delete_instance(command, self.state.instances)
                            transition_events = get_transition_events(
-                                self.state.instances, placement, self.state.tasks
+                                self.state.instances, placement
                            )
                            for cmd in cancel_unnecessary_downloads(
                                placement, self.state.downloads
@@ -301,7 +299,7 @@ class Master:
                                self.state.node_network,
                            )
                            transition_events = get_transition_events(
-                                self.state.instances, placement, self.state.tasks
+                                self.state.instances, placement
                            )
                            generated_events.extend(transition_events)
                        case CreateInstance():
@@ -311,7 +309,7 @@ class Master:
                                self.state.instances,
                            )
                            transition_events = get_transition_events(
-                                self.state.instances, placement, self.state.tasks
+                                self.state.instances, placement
                            )
                            generated_events.extend(transition_events)
                        case SendInputChunk(chunk=chunk):
@@ -321,18 +319,6 @@ class Master:
                                    chunk=chunk,
                                )
                            )
-                        case TaskCancelled():
-                            if (
-                                task_id := self.command_task_mapping.get(
-                                    command.cancelled_command_id
-                                )
-                            ) is not None:
-                                generated_events.append(
-                                    TaskStatusUpdated(
-                                        task_status=TaskStatus.Cancelled,
-                                        task_id=task_id,
-                                    )
-                                )
                        case TaskFinished():
                            generated_events.append(
                                TaskDeleted(
@@ -341,9 +327,10 @@ class Master:
                                    ]
                                )
                            )
-                            self.command_task_mapping.pop(
-                                command.finished_command_id, None
-                            )
+                            if command.finished_command_id in self.command_task_mapping:
+                                del self.command_task_mapping[
+                                    command.finished_command_id
+                                ]
                        case RequestEventLog():
                            # We should just be able to send everything, since other buffers will ignore old messages
                            # rate limit to 1000 at a time
--- a/src/exo/master/placement.py
+++ b/src/exo/master/placement.py
@@ -22,15 +22,9 @@ from exo.shared.types.commands import (
    PlaceInstance,
 )
 from exo.shared.types.common import NodeId
-from exo.shared.types.events import (
-    Event,
-    InstanceCreated,
-    InstanceDeleted,
-    TaskStatusUpdated,
-)
+from exo.shared.types.events import Event, InstanceCreated, InstanceDeleted
 from exo.shared.types.memory import Memory
 from exo.shared.types.profiling import MemoryUsage, NodeNetworkInfo
-from exo.shared.types.tasks import Task, TaskId, TaskStatus
 from exo.shared.types.worker.downloads import (
    DownloadOngoing,
    DownloadProgress,
@@ -192,7 +186,6 @@ def delete_instance(
 def get_transition_events(
    current_instances: Mapping[InstanceId, Instance],
    target_instances: Mapping[InstanceId, Instance],
-    tasks: Mapping[TaskId, Task],
 ) -> Sequence[Event]:
    events: list[Event] = []

@@ -208,18 +201,6 @@ def get_transition_events(
    # find instances to delete
    for instance_id in current_instances:
        if instance_id not in target_instances:
-            for task in tasks.values():
-                if task.instance_id == instance_id and task.task_status in [
-                    TaskStatus.Pending,
-                    TaskStatus.Running,
-                ]:
-                    events.append(
-                        TaskStatusUpdated(
-                            task_status=TaskStatus.Cancelled,
-                            task_id=task.task_id,
-                        )
-                    )
-
            events.append(
                InstanceDeleted(
                    instance_id=instance_id,
--- a/src/exo/master/tests/test_claude_tool_use.py
+++ b/src/exo/master/tests/test_claude_tool_use.py
@@ -4,11 +4,7 @@ import json
 from collections.abc import AsyncGenerator
 from typing import Any, cast

-from exo.master.adapters.claude import (
-    ClaudeMessagesResponse,
-    collect_claude_response,
-    generate_claude_stream,
-)
+from exo.master.adapters.claude import collect_claude_response, generate_claude_stream
 from exo.shared.types.api import ToolCallItem
 from exo.shared.types.chunks import ErrorChunk, TokenChunk, ToolCallChunk
 from exo.shared.types.common import CommandId, ModelId
@@ -21,18 +17,6 @@ async def _chunks_to_stream(
        yield chunk


-async def _collect_response(
-    command_id: CommandId,
-    model: str,
-    chunk_stream: AsyncGenerator[ErrorChunk | ToolCallChunk | TokenChunk, None],
-) -> ClaudeMessagesResponse:
-    """Helper to consume the async generator and parse the JSON response."""
-    parts: list[str] = []
-    async for part in collect_claude_response(command_id, model, chunk_stream):
-        parts.append(part)
-    return ClaudeMessagesResponse.model_validate_json("".join(parts))
-
-
 MODEL = ModelId("test-model")
 COMMAND_ID = CommandId("cmd_test123")

@@ -63,7 +47,7 @@ class TestCollectClaudeResponseToolUse:
                ],
            ),
        ]
-        response = await _collect_response(
+        response = await collect_claude_response(
            COMMAND_ID, "test-model", _chunks_to_stream(chunks)
        )

@@ -93,7 +77,7 @@ class TestCollectClaudeResponseToolUse:
                ],
            ),
        ]
-        response = await _collect_response(
+        response = await collect_claude_response(
            COMMAND_ID, "test-model", _chunks_to_stream(chunks)
        )

@@ -118,7 +102,7 @@ class TestCollectClaudeResponseToolUse:
                ],
            ),
        ]
-        response = await _collect_response(
+        response = await collect_claude_response(
            COMMAND_ID, "test-model", _chunks_to_stream(chunks)
        )

@@ -132,7 +116,7 @@ class TestCollectClaudeResponseToolUse:

    async def test_no_content_produces_empty_text_block(self):
        chunks: list[ErrorChunk | ToolCallChunk | TokenChunk] = []
-        response = await _collect_response(
+        response = await collect_claude_response(
            COMMAND_ID, "test-model", _chunks_to_stream(chunks)
        )
        assert len(response.content) == 1
--- a/src/exo/master/tests/test_master.py
+++ b/src/exo/master/tests/test_master.py
@@ -42,7 +42,7 @@ from exo.utils.channels import channel
@pytest.mark.asyncio
 async def test_master():
    keypair = get_node_id_keypair()
-    node_id = NodeId(keypair.to_string())
+    node_id = NodeId(keypair.to_peer_id().to_base58())
    session_id = SessionId(master_node_id=node_id, election_clock=0)

    ge_sender, global_event_receiver = channel[ForwarderEvent]()
@@ -75,7 +75,7 @@ async def test_master():
    async with anyio.create_task_group() as tg:
        tg.start_soon(master.run)

-        sender_node_id = NodeId(f"{keypair.to_string()}_sender")
+        sender_node_id = NodeId(f"{keypair.to_peer_id().to_base58()}_sender")
        # inject a NodeGatheredInfo event
        logger.info("inject a NodeGatheredInfo event")
        await local_event_sender.send(
--- a/src/exo/master/tests/test_placement.py
+++ b/src/exo/master/tests/test_placement.py
@@ -239,7 +239,7 @@ def test_get_transition_events_no_change(instance: Instance):
    target_instances = {instance_id: instance}

    # act
-    events = get_transition_events(current_instances, target_instances, {})
+    events = get_transition_events(current_instances, target_instances)

    # assert
    assert len(events) == 0
@@ -252,7 +252,7 @@ def test_get_transition_events_create_instance(instance: Instance):
    target_instances: dict[InstanceId, Instance] = {instance_id: instance}

    # act
-    events = get_transition_events(current_instances, target_instances, {})
+    events = get_transition_events(current_instances, target_instances)

    # assert
    assert len(events) == 1
@@ -266,7 +266,7 @@ def test_get_transition_events_delete_instance(instance: Instance):
    target_instances: dict[InstanceId, Instance] = {}

    # act
-    events = get_transition_events(current_instances, target_instances, {})
+    events = get_transition_events(current_instances, target_instances)

    # assert
    assert len(events) == 1
--- a/src/exo/routing/connection_message.py
+++ b/src/exo/routing/connection_message.py
@@ -0,0 +1,37 @@
+from enum import Enum
+
+from exo_pyo3_bindings import ConnectionUpdate, ConnectionUpdateType
+
+from exo.shared.types.common import NodeId
+from exo.utils.pydantic_ext import CamelCaseModel
+
+"""Serialisable types for Connection Updates/Messages"""
+
+
+class ConnectionMessageType(Enum):
+    Connected = 0
+    Disconnected = 1
+
+    @staticmethod
+    def from_update_type(update_type: ConnectionUpdateType):
+        match update_type:
+            case ConnectionUpdateType.Connected:
+                return ConnectionMessageType.Connected
+            case ConnectionUpdateType.Disconnected:
+                return ConnectionMessageType.Disconnected
+
+
+class ConnectionMessage(CamelCaseModel):
+    node_id: NodeId
+    connection_type: ConnectionMessageType
+    remote_ipv4: str
+    remote_tcp_port: int
+
+    @classmethod
+    def from_update(cls, update: ConnectionUpdate) -> "ConnectionMessage":
+        return cls(
+            node_id=NodeId(update.peer_id.to_base58()),
+            connection_type=ConnectionMessageType.from_update_type(update.update_type),
+            remote_ipv4=update.remote_ipv4,
+            remote_tcp_port=update.remote_tcp_port,
+        )
--- a/src/exo/routing/router.py
+++ b/src/exo/routing/router.py
@@ -16,19 +16,17 @@ from anyio.abc import TaskGroup
 from exo_pyo3_bindings import (
    AllQueuesFullError,
    Keypair,
+    NetworkingHandle,
    NoPeersSubscribedToTopicError,
-    PyMessage,
-    PySwarm,
 )
 from filelock import FileLock
 from loguru import logger

 from exo.shared.constants import EXO_NODE_ID_KEYPAIR
-from exo.shared.election import ConnectionMessage
-from exo.shared.types.common import NodeId
 from exo.utils.channels import Receiver, Sender, channel
 from exo.utils.pydantic_ext import CamelCaseModel

+from .connection_message import ConnectionMessage
 from .topics import CONNECTION_MESSAGES, PublishPolicy, TypedTopic


@@ -104,13 +102,13 @@ class TopicRouter[T: CamelCaseModel]:
 class Router:
    @classmethod
    def create(cls, identity: Keypair) -> "Router":
-        return cls(handle=PySwarm(identity))
+        return cls(handle=NetworkingHandle(identity))

-    def __init__(self, handle: PySwarm):
+    def __init__(self, handle: NetworkingHandle):
        self.topic_routers: dict[str, TopicRouter[CamelCaseModel]] = {}
        send, recv = channel[tuple[str, bytes]]()
        self.networking_receiver: Receiver[tuple[str, bytes]] = recv
-        self._net = handle
+        self._net: NetworkingHandle = handle
        self._tmp_networking_sender: Sender[tuple[str, bytes]] | None = send
        self._id_count = count()
        self._tg: TaskGroup | None = None
@@ -156,6 +154,7 @@ class Router:
                    router = self.topic_routers[topic]
                    tg.start_soon(router.run)
                tg.start_soon(self._networking_recv)
+                tg.start_soon(self._networking_recv_connection_messages)
                tg.start_soon(self._networking_publish)
                # Router only shuts down if you cancel it.
                await sleep_forever()
@@ -180,44 +179,38 @@ class Router:

    async def _networking_recv(self):
        while True:
-            try:
-                msg = await self._net.recv()
-            except NoPeersSubscribedToTopicError:
-                continue
-            except AllQueuesFullError:
-                logger.warning("All peer queues full, messages have been lost")
+            topic, data = await self._net.gossipsub_recv()
+            logger.trace(f"Received message on {topic} with payload {data}")
+            if topic not in self.topic_routers:
+                logger.warning(f"Received message on unknown or inactive topic {topic}")
                continue

-            match msg:
-                case PyMessage.Connection():
-                    if CONNECTION_MESSAGES.topic in self.topic_routers:
-                        router = self.topic_routers[CONNECTION_MESSAGES.topic]
-                        assert router.topic.model_type == ConnectionMessage
-                        router = cast(TopicRouter[ConnectionMessage], router)
-                        await router.publish(
-                            ConnectionMessage(
-                                node_id=NodeId(msg.node_id), connected=msg.connected
-                            )
-                        )
-                case PyMessage.Gossip():
-                    if msg.topic not in self.topic_routers:
-                        logger.warning(
-                            f"Received message on unknown or inactive topic {msg.topic}"
-                        )
-                        continue
-                    logger.trace(
-                        f"Received message on {msg.topic} with payload {msg.data}"
-                    )
-                    router = self.topic_routers[msg.topic]
-                    await router.publish_bytes(msg.data)
-                case _:
-                    raise ValueError("net recv returned something impossible")
+            router = self.topic_routers[topic]
+            await router.publish_bytes(data)
+
+    async def _networking_recv_connection_messages(self):
+        while True:
+            update = await self._net.connection_update_recv()
+            message = ConnectionMessage.from_update(update)
+            logger.trace(
+                f"Received message on connection_messages with payload {message}"
+            )
+            if CONNECTION_MESSAGES.topic in self.topic_routers:
+                router = self.topic_routers[CONNECTION_MESSAGES.topic]
+                assert router.topic.model_type == ConnectionMessage
+                router = cast(TopicRouter[ConnectionMessage], router)
+                await router.publish(message)

    async def _networking_publish(self):
        with self.networking_receiver as networked_items:
            async for topic, data in networked_items:
-                logger.trace(f"Sending message on {topic} with payload {data}")
-                await self._net.gossipsub_publish(topic, data)
+                try:
+                    logger.trace(f"Sending message on {topic} with payload {data}")
+                    await self._net.gossipsub_publish(topic, data)
+                except NoPeersSubscribedToTopicError:
+                    pass
+                except AllQueuesFullError:
+                    logger.warning(f"All peer queues full, dropping message on {topic}")


 def get_node_id_keypair(
@@ -228,7 +221,7 @@ def get_node_id_keypair(
    Obtain the :class:`PeerId` by from it.
    """
    # TODO(evan): bring back node id persistence once we figure out how to deal with duplicates
-    return Keypair.generate()
+    return Keypair.generate_ed25519()

    def lock_path(path: str | bytes | PathLike[str] | PathLike[bytes]) -> Path:
        return Path(str(path) + ".lock")
@@ -242,12 +235,12 @@ def get_node_id_keypair(
                protobuf_encoded = f.read()

                try:  # if decoded successfully, save & return
-                    return Keypair.deserialize(protobuf_encoded)
+                    return Keypair.from_protobuf_encoding(protobuf_encoded)
                except ValueError as e:  # on runtime error, assume corrupt file
                    logger.warning(f"Encountered error when trying to get keypair: {e}")

        # if no valid credentials, create new ones and persist
        with open(path, "w+b") as f:
-            keypair = Keypair.generate()
-            f.write(keypair.serialize())
+            keypair = Keypair.generate_ed25519()
+            f.write(keypair.to_protobuf_encoding())
            return keypair
--- a/src/exo/routing/topics.py
+++ b/src/exo/routing/topics.py
@@ -1,7 +1,8 @@
 from dataclasses import dataclass
 from enum import Enum

-from exo.shared.election import ConnectionMessage, ElectionMessage
+from exo.routing.connection_message import ConnectionMessage
+from exo.shared.election import ElectionMessage
 from exo.shared.types.commands import ForwarderCommand, ForwarderDownloadCommand
 from exo.shared.types.events import (
    ForwarderEvent,
--- a/src/exo/shared/apply.py
+++ b/src/exo/shared/apply.py
@@ -15,6 +15,7 @@ from exo.shared.types.events import (
    NodeDownloadProgress,
    NodeGatheredInfo,
    NodeTimedOut,
+    PrefillProgress,
    RunnerDeleted,
    RunnerStatusUpdated,
    TaskAcknowledged,
@@ -64,6 +65,7 @@ def event_apply(event: Event, state: State) -> State:
            | ChunkGenerated()
            | TaskAcknowledged()
            | InputChunkReceived()
+            | PrefillProgress()
            | TracesCollected()
            | TracesMerged()
        ):  # Pass-through events that don't modify state
@@ -218,6 +220,11 @@ def apply_node_timed_out(event: NodeTimedOut, state: State) -> State:
        key: value for key, value in state.downloads.items() if key != event.node_id
    }
    # Clean up all granular node mappings
+    node_identities = {
+        key: value
+        for key, value in state.node_identities.items()
+        if key != event.node_id
+    }
    node_memory = {
        key: value for key, value in state.node_memory.items() if key != event.node_id
    }
@@ -258,6 +265,7 @@ def apply_node_timed_out(event: NodeTimedOut, state: State) -> State:
            "downloads": downloads,
            "topology": topology,
            "last_seen": last_seen,
+            "node_identities": node_identities,
            "node_memory": node_memory,
            "node_disk": node_disk,
            "node_system": node_system,
--- a/src/exo/shared/election.py
+++ b/src/exo/shared/election.py
@@ -10,6 +10,7 @@ from anyio import (
 from anyio.abc import TaskGroup
 from loguru import logger

+from exo.routing.connection_message import ConnectionMessage
 from exo.shared.types.commands import ForwarderCommand
 from exo.shared.types.common import NodeId, SessionId
 from exo.utils.channels import Receiver, Sender
@@ -18,11 +19,6 @@ from exo.utils.pydantic_ext import CamelCaseModel
 DEFAULT_ELECTION_TIMEOUT = 3.0


-class ConnectionMessage(CamelCaseModel):
-    node_id: NodeId
-    connected: bool
-
-
 class ElectionMessage(CamelCaseModel):
    clock: int
    seniority: int
--- a/src/exo/shared/tests/test_election.py
+++ b/src/exo/shared/tests/test_election.py
@@ -1,7 +1,7 @@
 import pytest
 from anyio import create_task_group, fail_after, move_on_after

-from exo.routing.router import ConnectionMessage
+from exo.routing.connection_message import ConnectionMessage, ConnectionMessageType
 from exo.shared.election import Election, ElectionMessage, ElectionResult
 from exo.shared.types.commands import ForwarderCommand, TestCommand
 from exo.shared.types.common import NodeId, SessionId
@@ -330,7 +330,9 @@ async def test_connection_message_triggers_new_round_broadcast() -> None:
            await cm_tx.send(
                ConnectionMessage(
                    node_id=NodeId(),
-                    connected=True,
+                    connection_type=ConnectionMessageType.Connected,
+                    remote_ipv4="",
+                    remote_tcp_port=0,
                )
            )

--- a/src/exo/shared/tests/test_node_id_persistence.py
+++ b/src/exo/shared/tests/test_node_id_persistence.py
@@ -23,7 +23,7 @@ def _get_keypair_concurrent_subprocess_task(
    sem.release()
    # wait to be told to begin simultaneous read
    ev.wait()
-    queue.put(get_node_id_keypair().serialize())
+    queue.put(get_node_id_keypair().to_protobuf_encoding())


 def _get_keypair_concurrent(num_procs: int) -> bytes:
--- a/src/exo/shared/types/api.py
+++ b/src/exo/shared/types/api.py
@@ -3,7 +3,8 @@ from collections.abc import Generator
 from typing import Annotated, Any, Literal
 from uuid import uuid4

-from pydantic import BaseModel, Field
+from pydantic import BaseModel, Field, field_validator
+from pydantic_core import PydanticUseDefault

 from exo.shared.models.model_cards import ModelCard, ModelId
 from exo.shared.types.common import CommandId, NodeId
@@ -227,6 +228,13 @@ class PlaceInstanceParams(BaseModel):
    instance_meta: InstanceMeta = InstanceMeta.MlxRing
    min_nodes: int = 1

+    @field_validator("sharding", "instance_meta", mode="plain")
+    @classmethod
+    def use_default(cls, v: object):
+        if not v or not isinstance(v, (Sharding, InstanceMeta)):
+            raise PydanticUseDefault()
+        return v
+

 class CreateInstanceParams(BaseModel):
    instance: Instance
--- a/src/exo/shared/types/chunks.py
+++ b/src/exo/shared/types/chunks.py
@@ -77,3 +77,13 @@ class InputImageChunk(BaseChunk):


 GenerationChunk = TokenChunk | ImageChunk | ToolCallChunk | ErrorChunk
+
+
+class PrefillProgressData(TaggedModel):
+    """Data class for prefill progress events during streaming."""
+
+    processed_tokens: int
+    total_tokens: int
+
+
+StreamEvent = TokenChunk | PrefillProgressData
--- a/src/exo/shared/types/commands.py
+++ b/src/exo/shared/types/commands.py
@@ -48,10 +48,6 @@ class DeleteInstance(BaseCommand):
    instance_id: InstanceId


-class TaskCancelled(BaseCommand):
-    cancelled_command_id: CommandId
-
-
 class TaskFinished(BaseCommand):
    finished_command_id: CommandId

@@ -93,7 +89,6 @@ Command = (
    | PlaceInstance
    | CreateInstance
    | DeleteInstance
-    | TaskCancelled
    | TaskFinished
    | SendInputChunk
 )
--- a/src/exo/shared/types/events.py
+++ b/src/exo/shared/types/events.py
@@ -102,6 +102,12 @@ class InputChunkReceived(BaseEvent):
    chunk: InputImageChunk


+class PrefillProgress(BaseEvent):
+    command_id: CommandId
+    processed_tokens: int
+    total_tokens: int
+
+
 class TopologyEdgeCreated(BaseEvent):
    conn: Connection

@@ -148,6 +154,7 @@ Event = (
    | NodeDownloadProgress
    | ChunkGenerated
    | InputChunkReceived
+    | PrefillProgress
    | TopologyEdgeCreated
    | TopologyEdgeDeleted
    | TracesCollected
--- a/src/exo/shared/types/tasks.py
+++ b/src/exo/shared/types/tasks.py
@@ -24,7 +24,6 @@ class TaskStatus(str, Enum):
    Complete = "Complete"
    TimedOut = "TimedOut"
    Failed = "Failed"
-    Cancelled = "Cancelled"


 class BaseTask(TaggedModel):
@@ -61,11 +60,6 @@ class TextGeneration(BaseTask):  # emitted by Master
    error_message: str | None = Field(default=None)


-class CancelTask(BaseTask):
-    cancelled_task_id: TaskId
-    runner_id: RunnerId
-
-
 class ImageGeneration(BaseTask):  # emitted by Master
    command_id: CommandId
    task_params: ImageGenerationTaskParams
@@ -93,7 +87,6 @@ Task = (
    | LoadModel
    | StartWarmup
    | TextGeneration
-    | CancelTask
    | ImageGeneration
    | ImageEdits
    | Shutdown
--- a/src/exo/shared/types/worker/downloads.py
+++ b/src/exo/shared/types/worker/downloads.py
@@ -26,7 +26,6 @@ class DownloadProgressData(CamelCaseModel):
 class BaseDownloadProgress(TaggedModel):
    node_id: NodeId
    shard_metadata: ShardMetadata
-    model_directory: str = ""


 class DownloadPending(BaseDownloadProgress):
--- a/src/exo/shared/types/worker/runner_response.py
+++ b/src/exo/shared/types/worker/runner_response.py
@@ -62,8 +62,12 @@ class PartialImageResponse(BaseRunnerResponse):
 class ToolCallResponse(BaseRunnerResponse):
    tool_calls: list[ToolCallItem]
    usage: Usage | None
-    stats: GenerationStats | None = None


 class FinishedResponse(BaseRunnerResponse):
    pass
+
+
+class PrefillProgressResponse(BaseRunnerResponse):
+    processed_tokens: int
+    total_tokens: int
--- a/src/exo/utils/banner.py
+++ b/src/exo/utils/banner.py
@@ -1,7 +1,5 @@
-import sys
-
-
 def print_startup_banner(port: int) -> None:
+    """Print a prominent startup banner with API endpoint information."""
    dashboard_url = f"http://localhost:{port}"
    banner = f"""
 ╔═══════════════════════════════════════════════════════════════════════╗
@@ -29,4 +27,4 @@ def print_startup_banner(port: int) -> None:

 """

-    print(banner, file=sys.stderr)
+    print(banner)
--- a/src/exo/utils/channels.py
+++ b/src/exo/utils/channels.py
@@ -125,9 +125,7 @@ class MpSender[T]:
            self._state.buffer.put(item, block=True)

    async def send_async(self, item: T) -> None:
-        await to_thread.run_sync(
-            self.send, item, limiter=CapacityLimiter(1), abandon_on_cancel=True
-        )
+        await to_thread.run_sync(self.send, item, limiter=CapacityLimiter(1))

    def close(self) -> None:
        if not self._state.closed.is_set():
--- a/src/exo/worker/engines/mlx/generator/generate.py
+++ b/src/exo/worker/engines/mlx/generator/generate.py
@@ -160,7 +160,7 @@ def warmup_inference(
        max_tokens=50,
        sampler=sampler,
        prompt_cache=cache,
-        prefill_step_size=2048,
+        prefill_step_size=1024,
        kv_group_size=KV_GROUP_SIZE,
        kv_bits=KV_BITS,
    ):
@@ -252,6 +252,7 @@ def mlx_generate(
    task: TextGenerationTaskParams,
    prompt: str,
    kv_prefix_cache: KVPrefixCache | None = None,
+    on_prefill_progress: Callable[[int, int], None] | None = None,
    group: mx.distributed.Group | None = None,
 ) -> Generator[GenerationResponse]:
    # Ensure that generation stats only contains peak memory for this generation
@@ -306,7 +307,7 @@ def mlx_generate(
    max_stop_len = max((len(s) for s in stop_sequences), default=0)

    mx_barrier(group)
-    logger.info("Starting prefill")
+    logger.info("Ready to prefill")

    # Prefill cache with all tokens except the last one
    prefill_tps, prefill_tokens, ssm_snapshots_list = prefill(
@@ -345,6 +346,7 @@ def mlx_generate(
            prefill_step_size=1,
            kv_group_size=KV_GROUP_SIZE,
            kv_bits=KV_BITS,
+            prompt_progress_callback=on_prefill_progress,
        ),
        start=1,
    ):
@@ -393,11 +395,10 @@ def mlx_generate(
                    f"Model generated unexpected finish_reason: {out.finish_reason}"
                )

-            total_prompt_tokens = len(all_prompt_tokens)
            usage = Usage(
-                prompt_tokens=total_prompt_tokens,
+                prompt_tokens=int(out.prompt_tokens),
                completion_tokens=completion_tokens,
-                total_tokens=total_prompt_tokens + completion_tokens,
+                total_tokens=int(out.prompt_tokens) + completion_tokens,
                prompt_tokens_details=PromptTokensDetails(
                    cached_tokens=prefix_hit_length
                ),
--- a/src/exo/worker/engines/mlx/utils_mlx.py
+++ b/src/exo/worker/engines/mlx/utils_mlx.py
@@ -64,6 +64,8 @@ from exo.worker.runner.bootstrap import logger
 Group = mx.distributed.Group


+# TODO: Test this
+#  ALSO https://github.com/exo-explore/exo/pull/233#discussion_r2549683673
 def get_weights_size(model_shard_meta: ShardMetadata) -> Memory:
    return Memory.from_float_kb(
        (model_shard_meta.end_layer - model_shard_meta.start_layer)
@@ -81,6 +83,30 @@ class ModelLoadingTimeoutError(Exception):
    pass


+def mx_barrier(group: Group | None = None):
+    mx.eval(
+        mx.distributed.all_sum(
+            mx.array(1.0),
+            stream=mx.default_stream(mx.Device(mx.cpu)),
+            group=group,
+        )
+    )
+
+
+def broadcast_from_zero(value: int, group: Group | None = None):
+    if group is None:
+        return value
+
+    if group.rank() == 0:
+        a = mx.array([value], dtype=mx.int32)
+    else:
+        a = mx.array([0], dtype=mx.int32)
+
+    m = mx.distributed.all_sum(a, stream=mx.Device(mx.DeviceType.cpu), group=group)
+    mx.eval(m)
+    return int(m.item())
+
+
 class HostList(RootModel[list[str]]):
    @classmethod
    def from_hosts(cls, hosts: list[Host]) -> "HostList":
@@ -353,13 +379,7 @@ def load_tokenizer_for_model_id(
            return list(hf_tokenizer.model.encode(text, allowed_special="all"))  # pyright: ignore[reportUnknownMemberType,reportUnknownArgumentType]

        hf_tokenizer.encode = _patched_encode
-        return TokenizerWrapper(
-            hf_tokenizer,
-            eos_token_ids=eos_token_ids,
-            tool_call_start="<|tool_calls_section_begin|>",
-            tool_call_end="<|tool_calls_section_end|>",
-            tool_parser=_parse_kimi_tool_calls,
-        )
+        return TokenizerWrapper(hf_tokenizer, eos_token_ids=eos_token_ids)

    tokenizer = load_tokenizer(
        model_path,
@@ -571,61 +591,3 @@ def mlx_cleanup(
    import gc

    gc.collect()
-
-
-def mx_any(bool_: bool, group: Group | None) -> bool:
-    if group is None:
-        return bool_
-    num_true = mx.distributed.all_sum(
-        mx.array(bool_), group=group, stream=mx.default_stream(mx.Device(mx.cpu))
-    )
-    mx.eval(num_true)
-    return num_true.item() > 0
-
-
-def mx_barrier(group: Group | None):
-    if group is None:
-        return
-    mx.eval(
-        mx.distributed.all_sum(
-            mx.array(1.0), group=group, stream=mx.default_stream(mx.Device(mx.cpu))
-        )
-    )
-
-
-def _parse_kimi_tool_calls(text: str):
-    import regex as re
-
-    # kimi has a fixed function naming scheme, with a json formatted arg
-    #   functions.multiply:0<|tool_call_argument_begin|>{"a": 2, "b": 3}
-    _func_name_regex = re.compile(
-        r"^\s*((?:functions\.)?(.+?):\d+)\s*<\|tool_call_argument_begin\|>", re.DOTALL
-    )
-    _func_arg_regex = re.compile(r"<\|tool_call_argument_begin\|>\s*(.*)\s*", re.DOTALL)
-    _tool_call_split_regex = re.compile(
-        r"<\|tool_call_begin\|>(.*?)<\|tool_call_end\|>", re.DOTALL
-    )
-
-    def _parse_single_tool(text: str) -> dict[str, Any]:
-        func_name_match = _func_name_regex.search(text)
-        if func_name_match is None:
-            raise ValueError("No tool call found.")
-        tool_call_id = func_name_match.group(1)  # e.g. "functions.get_weather:0"
-        func_name = func_name_match.group(2)  # e.g. "get_weather"
-
-        func_args_match = _func_arg_regex.search(text)
-        if func_args_match is None:
-            raise ValueError("No tool call arguments found.")
-        func_args = func_args_match.group(1)
-        try:
-            arg_dct = json.loads(func_args)  # pyright: ignore[reportAny]
-        except Exception:
-            arg_dct = None
-
-        return dict(id=tool_call_id, name=func_name, arguments=arg_dct)
-
-    tool_matches = _tool_call_split_regex.findall(text)
-    if tool_matches:
-        return [_parse_single_tool(match) for match in tool_matches]  # pyright: ignore[reportAny]
-    else:
-        return [_parse_single_tool(text)]
--- a/src/exo/worker/main.py
+++ b/src/exo/worker/main.py
@@ -33,7 +33,6 @@ from exo.shared.types.events import (
 from exo.shared.types.multiaddr import Multiaddr
 from exo.shared.types.state import State
 from exo.shared.types.tasks import (
-    CancelTask,
    CreateRunner,
    DownloadModel,
    ImageEdits,
@@ -225,22 +224,15 @@ class Worker:
                        )
                    )
                case Shutdown(runner_id=runner_id):
-                    runner = self.runners.pop(runner_id)
                    try:
                        with fail_after(3):
-                            await runner.start_task(task)
+                            await self.runners.pop(runner_id).start_task(task)
                    except TimeoutError:
                        await self.event_sender.send(
                            TaskStatusUpdated(
                                task_id=task.task_id, task_status=TaskStatus.TimedOut
                            )
                        )
-                    finally:
-                        runner.shutdown()
-                case CancelTask(
-                    cancelled_task_id=cancelled_task_id, runner_id=runner_id
-                ):
-                    await self.runners[runner_id].cancel_task(cancelled_task_id)
                case ImageEdits() if task.task_params.total_input_chunks > 0:
                    # Assemble image from chunks and inject into task
                    cmd_id = task.command_id
@@ -278,18 +270,18 @@ class Worker:
                        del self.input_chunk_buffer[cmd_id]
                    if cmd_id in self.input_chunk_counts:
                        del self.input_chunk_counts[cmd_id]
-                    await self._start_runner_task(modified_task)
+                    await self.runners[self._task_to_runner_id(task)].start_task(
+                        modified_task
+                    )
                case task:
-                    await self._start_runner_task(task)
+                    await self.runners[self._task_to_runner_id(task)].start_task(task)

    def shutdown(self):
        self._tg.cancel_scope.cancel()

-    async def _start_runner_task(self, task: Task):
-        if (instance := self.state.instances.get(task.instance_id)) is not None:
-            await self.runners[
-                instance.shard_assignments.node_to_runner[self.node_id]
-            ].start_task(task)
+    def _task_to_runner_id(self, task: Task):
+        instance = self.state.instances[task.instance_id]
+        return instance.shard_assignments.node_to_runner[self.node_id]

    async def _nack_request(self, since_idx: int) -> None:
        # We request all events after (and including) the missing index.
@@ -328,6 +320,8 @@ class Worker:
            for event in self.out_for_delivery.copy().values():
                await self.local_event_sender.send(event)

+    ## Op Executors
+
    def _create_supervisor(self, task: CreateRunner) -> RunnerSupervisor:
        """Creates and stores a new AssignedRunner with initial downloading status."""
        runner = RunnerSupervisor.create(
--- a/src/exo/worker/plan.py
+++ b/src/exo/worker/plan.py
@@ -4,7 +4,6 @@ from collections.abc import Mapping, Sequence

 from exo.shared.types.common import CommandId, NodeId
 from exo.shared.types.tasks import (
-    CancelTask,
    ConnectToGroup,
    CreateRunner,
    DownloadModel,
@@ -54,14 +53,13 @@ def plan(
 ) -> Task | None:
    # Python short circuiting OR logic should evaluate these sequentially.
    return (
-        _cancel_tasks(runners, tasks)
-        or _kill_runner(runners, all_runners, instances)
+        _kill_runner(runners, all_runners, instances)
        or _create_runner(node_id, runners, instances)
        or _model_needs_download(node_id, runners, global_download_status)
        or _init_distributed_backend(runners, all_runners)
        or _load_model(runners, all_runners, global_download_status)
        or _ready_to_warmup(runners, all_runners)
-        or _pending_tasks(runners, tasks, all_runners, input_chunk_buffer or {})
+        or _pending_tasks(runners, tasks, all_runners, input_chunk_buffer)
    )


@@ -272,7 +270,7 @@ def _pending_tasks(
    runners: Mapping[RunnerId, RunnerSupervisor],
    tasks: Mapping[TaskId, Task],
    all_runners: Mapping[RunnerId, RunnerStatus],
-    input_chunk_buffer: Mapping[CommandId, dict[int, str]],
+    input_chunk_buffer: Mapping[CommandId, dict[int, str]] | None = None,
 ) -> Task | None:
    for task in tasks.values():
        # for now, just forward chat completions
@@ -286,7 +284,7 @@ def _pending_tasks(
        if isinstance(task, ImageEdits) and task.task_params.total_input_chunks > 0:
            cmd_id = task.command_id
            expected = task.task_params.total_input_chunks
-            received = len(input_chunk_buffer.get(cmd_id, {}))
+            received = len((input_chunk_buffer or {}).get(cmd_id, {}))
            if received < expected:
                continue  # Wait for all chunks to arrive

@@ -294,33 +292,16 @@ def _pending_tasks(
            if task.instance_id != runner.bound_instance.instance.instance_id:
                continue

-            # the task status _should_ be set to completed by the LAST runner
-            # it is currently set by the first
-            # this is definitely a hack
+            # I have a design point here; this is a state race in disguise as the task status doesn't get updated to completed fast enough
+            # however, realistically the task status should be set to completed by the LAST runner, so this is a true race
+            # the actual solution is somewhat deeper than this bypass - TODO!
            if task.task_id in runner.completed:
                continue

+            # TODO: Check ordering aligns with MLX distributeds expectations.
+
            if isinstance(runner.status, RunnerReady) and all(
                isinstance(all_runners[global_runner_id], (RunnerReady, RunnerRunning))
                for global_runner_id in runner.bound_instance.instance.shard_assignments.runner_to_shard
            ):
                return task
-
-
-def _cancel_tasks(
-    runners: Mapping[RunnerId, RunnerSupervisor],
-    tasks: Mapping[TaskId, Task],
-) -> Task | None:
-    for task in tasks.values():
-        if task.task_status != TaskStatus.Cancelled:
-            continue
-        for runner_id, runner in runners.items():
-            if task.instance_id != runner.bound_instance.instance.instance_id:
-                continue
-            if task.task_id in runner.cancelled:
-                continue
-            return CancelTask(
-                instance_id=task.instance_id,
-                cancelled_task_id=task.task_id,
-                runner_id=runner_id,
-            )
--- a/src/exo/worker/runner/bootstrap.py
+++ b/src/exo/worker/runner/bootstrap.py
@@ -3,7 +3,7 @@ import os
 import loguru

 from exo.shared.types.events import Event, RunnerStatusUpdated
-from exo.shared.types.tasks import Task, TaskId
+from exo.shared.types.tasks import Task
 from exo.shared.types.worker.instances import BoundInstance, MlxJacclInstance
 from exo.shared.types.worker.runners import RunnerFailed
 from exo.utils.channels import ClosedResourceError, MpReceiver, MpSender
@@ -15,7 +15,6 @@ def entrypoint(
    bound_instance: BoundInstance,
    event_sender: MpSender[Event],
    task_receiver: MpReceiver[Task],
-    cancel_receiver: MpReceiver[TaskId],
    _logger: "loguru.Logger",
 ) -> None:
    fast_synch_override = os.environ.get("EXO_FAST_SYNCH")
@@ -39,7 +38,7 @@ def entrypoint(
    try:
        from exo.worker.runner.runner import main

-        main(bound_instance, event_sender, task_receiver, cancel_receiver)
+        main(bound_instance, event_sender, task_receiver)
    except ClosedResourceError:
        logger.warning("Runner communication closed unexpectedly")
    except Exception as e:
--- a/src/exo/worker/runner/runner.py
+++ b/src/exo/worker/runner/runner.py
@@ -1,10 +1,10 @@
 import base64
-import math
+import json
 import resource
 import time
 from collections.abc import Generator
 from functools import cache
-from typing import Literal
+from typing import Any, Callable, Literal

 import mlx.core as mx
 from mlx_lm.models.gpt_oss import Model as GptOssModel
@@ -15,6 +15,7 @@ from openai_harmony import (  # pyright: ignore[reportMissingTypeStubs]
    StreamableParser,
    load_harmony_encoding,
 )
+from pydantic import ValidationError

 from exo.shared.constants import EXO_MAX_CHUNK_SIZE, EXO_TRACING_ENABLED
 from exo.shared.models.model_cards import ModelId, ModelTask
@@ -25,6 +26,7 @@ from exo.shared.types.common import CommandId
 from exo.shared.types.events import (
    ChunkGenerated,
    Event,
+    PrefillProgress,
    RunnerStatusUpdated,
    TaskAcknowledged,
    TaskStatusUpdated,
@@ -87,12 +89,9 @@ from exo.worker.engines.mlx.utils_mlx import (
    initialize_mlx,
    load_mlx_items,
    mlx_force_oom,
-    mx_any,
 )
 from exo.worker.runner.bootstrap import logger

-from .tool_parsers import ToolParser, make_mlx_parser
-

 def _is_primary_output_node(shard_metadata: ShardMetadata) -> bool:
    """Check if this node is the primary output node for image generation.
@@ -114,7 +113,6 @@ def main(
    bound_instance: BoundInstance,
    event_sender: MpSender[Event],
    task_receiver: MpReceiver[Task],
-    cancel_receiver: MpReceiver[TaskId],
 ):
    soft, hard = resource.getrlimit(resource.RLIMIT_NOFILE)
    resource.setrlimit(resource.RLIMIT_NOFILE, (min(max(soft, 2048), hard), hard))
@@ -132,16 +130,11 @@ def main(
        time.sleep(timeout)

    setup_start_time = time.time()
-    cancelled_tasks = set[TaskId]()

-    # type checker was unhappy with me - splitting these fixed it
-    inference_model: Model | None = None
-    image_model: DistributedImageModel | None = None
+    model: Model | DistributedImageModel | None = None
    tokenizer = None
-    tool_parser: ToolParser | None = None
    group = None
    kv_prefix_cache: KVPrefixCache | None = None
-    check_for_cancel_every: int | None = None

    current_status: RunnerStatus = RunnerIdle()
    logger.info("runner created")
@@ -154,7 +147,6 @@ def main(
            if task.task_id in seen:
                logger.warning("repeat task - potential error")
            seen.add(task.task_id)
-            cancelled_tasks.discard(TaskId("CANCEL_CURRENT_TASK"))
            event_sender.send(
                TaskStatusUpdated(task_id=task.task_id, task_status=TaskStatus.Running)
            )
@@ -200,28 +192,19 @@ def main(
                        time.sleep(0.5)

                    if ModelTask.TextGeneration in shard_metadata.model_card.tasks:
-                        inference_model, tokenizer = load_mlx_items(
+                        model, tokenizer = load_mlx_items(
                            bound_instance, group, on_timeout=on_model_load_timeout
                        )
                        logger.info(
-                            f"model has_tool_calling={tokenizer.has_tool_calling} using tokens {tokenizer.tool_call_start}, {tokenizer.tool_call_end}"
+                            f"model has_tool_calling={tokenizer.has_tool_calling}"
                        )
-                        if tokenizer.has_tool_calling:
-                            assert tokenizer.tool_call_start
-                            assert tokenizer.tool_call_end
-                            assert tokenizer.tool_parser  # pyright: ignore[reportAny]
-                            tool_parser = make_mlx_parser(
-                                tokenizer.tool_call_start,
-                                tokenizer.tool_call_end,
-                                tokenizer.tool_parser,  # pyright: ignore[reportAny]
-                            )
                        kv_prefix_cache = KVPrefixCache(group)

                    elif (
                        ModelTask.TextToImage in shard_metadata.model_card.tasks
                        or ModelTask.ImageToImage in shard_metadata.model_card.tasks
                    ):
-                        image_model = initialize_image_model(bound_instance)
+                        model = initialize_image_model(bound_instance)
                    else:
                        raise ValueError(
                            f"Unknown model task(s): {shard_metadata.model_card.tasks}"
@@ -229,6 +212,8 @@ def main(
                    current_status = RunnerLoaded()
                    logger.info("runner loaded")
                case StartWarmup() if isinstance(current_status, RunnerLoaded):
+                    assert model
+
                    current_status = RunnerWarmingUp()
                    logger.info("runner warming up")
                    event_sender.send(
@@ -240,31 +225,16 @@ def main(

                    logger.info(f"warming up inference for instance: {instance}")
                    if ModelTask.TextGeneration in shard_metadata.model_card.tasks:
-                        assert inference_model
+                        assert not isinstance(model, DistributedImageModel)
                        assert tokenizer

-                        t = time.monotonic()
                        toks = warmup_inference(
-                            model=inference_model,
+                            model=model,
                            tokenizer=tokenizer,
                            group=group,
+                            # kv_prefix_cache=kv_prefix_cache,  # supply for warmup-time prefix caching
                        )
                        logger.info(f"warmed up by generating {toks} tokens")
-                        check_for_cancel_every = min(
-                            math.ceil(toks / min(time.monotonic() - t, 0.001)), 100
-                        )
-                        if group is not None:
-                            check_for_cancel_every = int(
-                                mx.max(
-                                    mx.distributed.all_gather(
-                                        mx.array([check_for_cancel_every]), group=group
-                                    )
-                                ).item()
-                            )
-
-                        logger.info(
-                            f"runner checking for cancellation every {check_for_cancel_every} tokens"
-                        )
                        logger.info(
                            f"runner initialized in {time.time() - setup_start_time} seconds"
                        )
@@ -272,8 +242,8 @@ def main(
                        ModelTask.TextToImage in shard_metadata.model_card.tasks
                        or ModelTask.ImageToImage in shard_metadata.model_card.tasks
                    ):
-                        assert image_model
-                        image = warmup_image_generator(model=image_model)
+                        assert isinstance(model, DistributedImageModel)
+                        image = warmup_image_generator(model=model)
                        if image is not None:
                            logger.info(f"warmed up by generating {image.size} image")
                        else:
@@ -293,9 +263,20 @@ def main(
                        )
                    )
                    event_sender.send(TaskAcknowledged(task_id=task.task_id))
-                    assert inference_model
+
+                    assert model and not isinstance(model, DistributedImageModel)
                    assert tokenizer
-                    assert check_for_cancel_every
+
+                    # Define callback to send prefill progress events directly
+                    def on_prefill_progress(processed: int, total: int) -> None:
+                        if device_rank == 0:
+                            event_sender.send(
+                                PrefillProgress(
+                                    command_id=command_id,
+                                    processed_tokens=processed,
+                                    total_tokens=total,
+                                )
+                            )

                    try:
                        _check_for_debug_prompts(task_params)
@@ -305,11 +286,12 @@ def main(

                        # Generate responses using the actual MLX generation
                        mlx_generator = mlx_generate(
-                            model=inference_model,
+                            model=model,
                            tokenizer=tokenizer,
                            task=task_params,
                            prompt=prompt,
                            kv_prefix_cache=kv_prefix_cache,
+                            on_prefill_progress=on_prefill_progress,
                            group=group,
                        )

@@ -320,25 +302,34 @@ def main(
                                mlx_generator, tokenizer
                            )

+                        # Kimi-K2 has tool call sections - we don't care about them
+                        if "kimi" in shard_metadata.model_card.model_id.lower():
+                            mlx_generator = filter_kimi_tokens(mlx_generator)
+                            patch_kimi_tokenizer(tokenizer)
+
+                        # GLM models need patched parser (upstream has bug with None regex match)
+                        elif "glm" in shard_metadata.model_card.model_id.lower():
+                            patch_glm_tokenizer(tokenizer)
+
                        # GPT-OSS specific parsing to match other model formats.
-                        if isinstance(inference_model, GptOssModel):
+                        elif isinstance(model, GptOssModel):
                            mlx_generator = parse_gpt_oss(mlx_generator)
-                        elif tool_parser:
-                            mlx_generator = parse_tool_calls(mlx_generator, tool_parser)
+
+                        if tokenizer.has_tool_calling and not isinstance(
+                            model, GptOssModel
+                        ):
+                            assert tokenizer.tool_call_start
+                            assert tokenizer.tool_call_end
+                            assert tokenizer.tool_parser  # pyright: ignore[reportAny]
+                            mlx_generator = parse_tool_calls(
+                                mlx_generator,
+                                tokenizer.tool_call_start,
+                                tokenizer.tool_call_end,
+                                tokenizer.tool_parser,  # pyright: ignore[reportAny]
+                            )

                        completion_tokens = 0
-                        tokens_since_last_cancel_check = 0
                        for response in mlx_generator:
-                            tokens_since_last_cancel_check += 1
-                            if tokens_since_last_cancel_check >= check_for_cancel_every:
-                                tokens_since_last_cancel_check = 0
-                                cancelled_tasks.update(cancel_receiver.collect())
-                                want_to_cancel = (task.task_id in cancelled_tasks) or (
-                                    TaskId("CANCEL_CURRENT_TASK") in cancelled_tasks
-                                )
-                                if mx_any(want_to_cancel, group):
-                                    break
-
                            match response:
                                case GenerationResponse():
                                    completion_tokens += 1
@@ -386,7 +377,6 @@ def main(
                                                    tool_calls=response.tool_calls,
                                                    model=shard_metadata.model_card.model_id,
                                                    usage=response.usage,
-                                                    stats=response.stats,
                                                ),
                                            )
                                        )
@@ -411,7 +401,7 @@ def main(
                case ImageGeneration(
                    task_params=task_params, command_id=command_id
                ) if isinstance(current_status, RunnerReady):
-                    assert image_model
+                    assert isinstance(model, DistributedImageModel)
                    logger.info(f"received image generation request: {str(task)[:500]}")
                    current_status = RunnerRunning()
                    logger.info("runner running")
@@ -424,9 +414,7 @@ def main(

                    try:
                        image_index = 0
-                        for response in generate_image(
-                            model=image_model, task=task_params
-                        ):
+                        for response in generate_image(model=model, task=task_params):
                            is_primary_output = _is_primary_output_node(shard_metadata)

                            if is_primary_output:
@@ -476,7 +464,7 @@ def main(
                case ImageEdits(task_params=task_params, command_id=command_id) if (
                    isinstance(current_status, RunnerReady)
                ):
-                    assert image_model
+                    assert isinstance(model, DistributedImageModel)
                    logger.info(f"received image edits request: {str(task)[:500]}")
                    current_status = RunnerRunning()
                    logger.info("runner running")
@@ -489,9 +477,7 @@ def main(

                    try:
                        image_index = 0
-                        for response in generate_image(
-                            model=image_model, task=task_params
-                        ):
+                        for response in generate_image(model=model, task=task_params):
                            if _is_primary_output_node(shard_metadata):
                                match response:
                                    case PartialImageResponse():
@@ -550,20 +536,14 @@ def main(
                    raise ValueError(
                        f"Received {task.__class__.__name__} outside of state machine in {current_status=}"
                    )
-            was_cancelled = (task.task_id in cancelled_tasks) or (
-                TaskId("CANCEL_CURRENT_TASK") in cancelled_tasks
+            event_sender.send(
+                TaskStatusUpdated(task_id=task.task_id, task_status=TaskStatus.Complete)
            )
-            if not was_cancelled:
-                event_sender.send(
-                    TaskStatusUpdated(
-                        task_id=task.task_id, task_status=TaskStatus.Complete
-                    )
-                )
            event_sender.send(
                RunnerStatusUpdated(runner_id=runner_id, runner_status=current_status)
            )
            if isinstance(current_status, RunnerShutdown):
-                del inference_model, image_model, tokenizer, group
+                del model, tokenizer, group
                mx.clear_cache()
                import gc

@@ -577,8 +557,21 @@ def get_gpt_oss_encoding():
    return encoding


+def filter_kimi_tokens(
+    responses: Generator[GenerationResponse | ToolCallResponse],
+) -> Generator[GenerationResponse]:
+    for resp in responses:
+        assert isinstance(resp, GenerationResponse)
+        if (
+            resp.text == "<|tool_calls_section_begin|>"
+            or resp.text == "<|tool_calls_section_end|>"
+        ):
+            continue
+        yield resp
+
+
 def parse_gpt_oss(
-    responses: Generator[GenerationResponse],
+    responses: Generator[GenerationResponse | ToolCallResponse],
 ) -> Generator[GenerationResponse | ToolCallResponse]:
    encoding = get_gpt_oss_encoding()
    stream = StreamableParser(encoding, role=Role.ASSISTANT)
@@ -635,9 +628,9 @@ def parse_gpt_oss(


 def parse_thinking_models(
-    responses: Generator[GenerationResponse],
+    responses: Generator[GenerationResponse | ToolCallResponse],
    tokenizer: TokenizerWrapper,
-) -> Generator[GenerationResponse]:
+) -> Generator[GenerationResponse | ToolCallResponse]:
    """
    For models that inject thinking tags in the prompt (like GLM-4.7),
    prepend the thinking tag to the output stream so the frontend
@@ -758,55 +751,218 @@ def _process_image_response(


 def parse_tool_calls(
-    responses: Generator[GenerationResponse], tool_parser: ToolParser
+    responses: Generator[GenerationResponse | ToolCallResponse],
+    tool_call_start: str,
+    tool_call_end: str,
+    tool_parser: Callable[[str], dict[str, Any] | list[dict[str, Any]]],
 ) -> Generator[GenerationResponse | ToolCallResponse]:
    in_tool_call = False
    tool_call_text_parts: list[str] = []
    for response in responses:
-        if response.text.startswith(tool_parser.start_parsing):
+        assert isinstance(response, GenerationResponse)
+        # assumption: the tool call start is one token
+        if response.text == tool_call_start:
            in_tool_call = True
-
-        if in_tool_call:
-            tool_call_text_parts.append(response.text)
-            if response.text.endswith(tool_parser.end_parsing):
-                # parse the actual tool calls from the tool call text
-                parsed = tool_parser.parse_tool_calls(
-                    "".join(tool_call_text_parts).strip()
-                )
+            continue
+        # assumption: the tool call end is one token
+        if in_tool_call and response.text == tool_call_end:
+            try:
+                # tool_parser returns an arbitrarily nested python dictionary
+                # we actually don't want the python dictionary, we just want to
+                # parse the top level { function: ..., arguments: ... } structure
+                # as we're just gonna hand it back to the api anyway
+                parsed = tool_parser("".join(tool_call_text_parts).strip())
                logger.info(f"parsed {tool_call_text_parts=} into {parsed=}")
-                if parsed is not None:
-                    yield ToolCallResponse(
-                        tool_calls=parsed, usage=response.usage, stats=response.stats
-                    )
+                if isinstance(parsed, list):
+                    tools = [_validate_single_tool(tool) for tool in parsed]
                else:
-                    logger.warning(
-                        f"tool call parsing failed for text {''.join(tool_call_text_parts)}"
-                    )
-                    response.text = "".join(tool_call_text_parts)
-                    yield response
+                    tools = [_validate_single_tool(parsed)]
+                yield ToolCallResponse(tool_calls=tools, usage=response.usage)

-                in_tool_call = False
-                tool_call_text_parts = []
-                continue
-
-            if response.finish_reason is not None:
-                logger.info(
-                    "tool call parsing interrupted, yield partial tool call as text"
-                )
-                response = response.model_copy(
-                    update={
-                        "text": "".join(tool_call_text_parts),
-                        "token": 0,
-                    }
+            except (
+                json.JSONDecodeError,
+                ValidationError,
+                ValueError,
+                AttributeError,
+            ) as e:
+                # ValueError: our parsers raise this for malformed tool calls
+                # AttributeError: upstream parsers (e.g. glm47) may raise this when regex doesn't match
+                logger.opt(exception=e).warning("tool call parsing failed")
+                # assumption: talking about tool calls, not making a tool call
+                response.text = (
+                    tool_call_start + "".join(tool_call_text_parts) + tool_call_end
                )
                yield response

+            in_tool_call = False
+            tool_call_text_parts = []
            continue

+        if in_tool_call:
+            tool_call_text_parts.append(response.text)
+            if response.finish_reason is not None:
+                logger.info(
+                    "toll call parsing interrupted, yield partial tool call as text"
+                )
+                yield GenerationResponse(
+                    text=tool_call_start + "".join(tool_call_text_parts),
+                    token=0,
+                    finish_reason=response.finish_reason,
+                    usage=None,
+                )
+            continue
        # fallthrough
        yield response


+def patch_kimi_tokenizer(tokenizer: TokenizerWrapper):
+    """
+    Version of to-be-upstreamed kimi-k2 tool parser
+    """
+    import ast
+    import json
+    from typing import Any
+
+    import regex as re
+
+    # kimi has a fixed function naming scheme, with a json formatted arg
+    #   functions.multiply:0 <|tool_call_argument_begin|> {"a": 2, "b": 3}
+    #   Also needs to handle tools like call_0<|tool_call_argument_begin|>{"filePath": "..."}
+    _func_name_regex = re.compile(
+        r"^\s*(.+)[:](\d+)\s*<\|tool_call_argument_begin\|>", re.DOTALL
+    )
+    _func_arg_regex = re.compile(r"<\|tool_call_argument_begin\|>\s*(.*)\s*", re.DOTALL)
+
+    # kimi has a tool_calls_section - we're leaving this up to the caller to handle
+    tool_call_start = "<|tool_call_begin|>"
+    tool_call_end = "<|tool_call_end|>"
+
+    def _deserialize(value: str) -> Any:  # pyright: ignore[reportAny]
+        try:
+            return json.loads(value)  # pyright: ignore[reportAny]
+        except Exception:
+            pass
+
+        try:
+            return ast.literal_eval(value)  # pyright: ignore[reportAny]
+        except Exception:
+            pass
+        return value
+
+    def parse_tool_call(text: str, tools: Any | None = None):
+        func_name_match = _func_name_regex.search(text)
+        if func_name_match is None:
+            raise ValueError(f"Could not parse function name from tool call: {text!r}")
+        original_func_name = func_name_match.group(1)
+        tool_id = func_name_match.group(2)
+        # strip off the `functions.` prefix, if it exists.
+        func_name = original_func_name[original_func_name.find(".") + 1 :]
+
+        func_args_match = _func_arg_regex.search(text)
+        if func_args_match is None:
+            raise ValueError(f"Could not parse function args from tool call: {text!r}")
+        func_args = func_args_match.group(1)
+        # the args should be valid json - no need to check against our tools to deserialize
+        arg_dct = _deserialize(func_args)  # pyright: ignore[reportAny]
+
+        return dict(
+            id=f"{original_func_name}:{tool_id}",
+            name=func_name,
+            arguments=arg_dct,  # pyright: ignore[reportAny]
+        )
+
+    tokenizer._tool_call_start = tool_call_start
+    tokenizer._tool_call_end = tool_call_end
+    tokenizer._tool_parser = parse_tool_call
+
+
+def patch_glm_tokenizer(tokenizer: TokenizerWrapper):
+    """
+    Fixed version of mlx_lm's glm47 tool parser that handles regex match failures.
+    """
+    import ast
+    import json
+    from typing import Any
+
+    import regex as re
+
+    _func_name_regex = re.compile(r"^(.*?)<arg_key>", re.DOTALL)
+    _func_arg_regex = re.compile(
+        r"<arg_key>(.*?)</arg_key>(?:\n|\s)*<arg_value>(.*?)(?:</arg_value>|(?=<arg_key>)|$)",
+        re.DOTALL,
+    )
+
+    tool_call_start = "<tool_call>"
+    tool_call_end = "</tool_call>"
+
+    def _is_string_type(
+        tool_name: str,
+        arg_name: str,
+        tools: list[Any] | None,
+    ) -> bool:
+        if tools is None:
+            return False
+        for tool in tools:  # pyright: ignore[reportAny]
+            func = tool["function"]  # pyright: ignore[reportAny]
+            if func["name"] == tool_name:
+                params = func["parameters"]  # pyright: ignore[reportAny]
+                if params is None:
+                    return False
+                props = params.get("properties", {})  # pyright: ignore[reportAny]
+                arg_props = props.get(arg_name, {})  # pyright: ignore[reportAny]
+                arg_type = arg_props.get("type", None)  # pyright: ignore[reportAny]
+                return arg_type == "string"  # pyright: ignore[reportAny]
+        return False
+
+    def _deserialize(value: str) -> Any:  # pyright: ignore[reportAny]
+        try:
+            return json.loads(value)  # pyright: ignore[reportAny]
+        except Exception:
+            pass
+        try:
+            return ast.literal_eval(value)  # pyright: ignore[reportAny]
+        except Exception:
+            pass
+        return value
+
+    def parse_tool_call(text: str, tools: list[Any] | None = None):
+        func_name_match = _func_name_regex.search(text)
+        if func_name_match is None:
+            raise ValueError(f"Could not parse function name from tool call: {text!r}")
+        func_name = func_name_match.group(1)
+
+        pairs = _func_arg_regex.findall(text)
+        arg_dct: dict[str, Any] = {}
+        for key, value in pairs:  # pyright: ignore[reportAny]
+            arg_key = key.strip()  # pyright: ignore[reportAny]
+            arg_val = value.strip()  # pyright: ignore[reportAny]
+            if not _is_string_type(func_name, arg_key, tools):  # pyright: ignore[reportAny]
+                arg_val = _deserialize(arg_val)  # pyright: ignore[reportAny]
+            arg_dct[arg_key] = arg_val
+        return dict(name=func_name, arguments=arg_dct)
+
+    tokenizer._tool_call_start = tool_call_start
+    tokenizer._tool_call_end = tool_call_end
+    tokenizer._tool_parser = parse_tool_call
+
+
+def _validate_single_tool(obj: dict[str, Any]) -> ToolCallItem:
+    if (
+        ((name := obj.get("name")) is not None)
+        and ((args := obj.get("arguments")) is not None)
+        and isinstance(name, str)
+    ):
+        raw_id: object = obj.get("id")
+        extra = {"id": str(raw_id)} if raw_id is not None else {}
+        return ToolCallItem(
+            **extra,
+            name=name,
+            arguments=json.dumps(args),
+        )
+    else:
+        raise ValidationError
+
+
 EXO_RUNNER_MUST_FAIL = "EXO RUNNER MUST FAIL"
 EXO_RUNNER_MUST_OOM = "EXO RUNNER MUST OOM"
 EXO_RUNNER_MUST_TIMEOUT = "EXO RUNNER MUST TIMEOUT"
--- a/src/exo/worker/runner/runner_supervisor.py
+++ b/src/exo/worker/runner/runner_supervisor.py
@@ -47,11 +47,9 @@ class RunnerSupervisor:
    _ev_recv: MpReceiver[Event]
    _task_sender: MpSender[Task]
    _event_sender: Sender[Event]
-    _cancel_sender: MpSender[TaskId]
    status: RunnerStatus = field(default_factory=RunnerIdle, init=False)
    pending: dict[TaskId, anyio.Event] = field(default_factory=dict, init=False)
    completed: set[TaskId] = field(default_factory=set, init=False)
-    cancelled: set[TaskId] = field(default_factory=set, init=False)

    @classmethod
    def create(
@@ -62,8 +60,8 @@ class RunnerSupervisor:
        initialize_timeout: float = 400,
    ) -> Self:
        ev_send, ev_recv = mp_channel[Event]()
+        # A task is kind of a runner command
        task_sender, task_recv = mp_channel[Task]()
-        cancel_sender, cancel_recv = mp_channel[TaskId]()

        runner_process = Process(
            target=entrypoint,
@@ -71,7 +69,6 @@ class RunnerSupervisor:
                bound_instance,
                ev_send,
                task_recv,
-                cancel_recv,
                logger,
            ),
            daemon=True,
@@ -86,7 +83,6 @@ class RunnerSupervisor:
            initialize_timeout=initialize_timeout,
            _ev_recv=ev_recv,
            _task_sender=task_sender,
-            _cancel_sender=cancel_sender,
            _event_sender=event_sender,
        )

@@ -101,8 +97,6 @@ class RunnerSupervisor:
        self._ev_recv.close()
        self._task_sender.close()
        self._event_sender.close()
-        self._cancel_sender.send(TaskId("CANCEL_CURRENT_TASK"))
-        self._cancel_sender.close()
        self.runner_process.join(1)
        if not self.runner_process.is_alive():
            logger.info("Runner process succesfully terminated")
@@ -118,6 +112,14 @@ class RunnerSupervisor:
        logger.critical("Runner process didn't respond to SIGTERM, killing")
        self.runner_process.kill()

+        self.runner_process.join(1)
+        if not self.runner_process.is_alive():
+            return
+
+        logger.critical(
+            "Runner process didn't respond to SIGKILL. System resources may have leaked"
+        )
+
    async def start_task(self, task: Task):
        if task.task_id in self.pending:
            logger.warning(
@@ -139,17 +141,6 @@ class RunnerSupervisor:
            return
        await event.wait()

-    async def cancel_task(self, task_id: TaskId):
-        if task_id in self.completed:
-            logger.info(f"Unable to cancel {task_id} as it has been completed")
-            return
-        self.cancelled.add(task_id)
-        with anyio.move_on_after(0.5) as scope:
-            await self._cancel_sender.send_async(task_id)
-        if scope.cancel_called:
-            logger.error("RunnerSupervisor cancel pipe blocked")
-            await self._check_runner(TimeoutError("cancel pipe blocked"))
-
    async def _forward_events(self):
        with self._ev_recv as events:
            try:
--- a/src/exo/worker/runner/tool_parsers.py
+++ b/src/exo/worker/runner/tool_parsers.py
@@ -1,72 +0,0 @@
-import json
-from dataclasses import dataclass
-from typing import Any, Callable
-
-from exo.shared.types.api import ToolCallItem
-
-
-@dataclass
-class ToolParser:
-    start_parsing: str
-    end_parsing: str
-    parse_tool_calls: Callable[[str], list[ToolCallItem] | None]
-
-
-def make_mlx_parser(
-    tool_call_start: str,
-    tool_call_end: str,
-    tool_parser: Callable[[str], dict[str, Any] | list[dict[str, Any]]],
-) -> ToolParser:
-    def parse_tool_calls(text: str) -> list[ToolCallItem] | None:
-        try:
-            text = text.removeprefix(tool_call_start)
-            text = text.removesuffix(tool_call_end)
-            parsed = tool_parser(text)
-            if isinstance(parsed, list):
-                return [ToolCallItem.model_validate(_flatten(p)) for p in parsed]
-            else:
-                return [ToolCallItem.model_validate(_flatten(parsed))]
-
-        except Exception:
-            return None
-
-    return ToolParser(
-        start_parsing=tool_call_start,
-        end_parsing=tool_call_end,
-        parse_tool_calls=parse_tool_calls,
-    )
-
-
-# TODO / example code:
-def _parse_json_calls(text: str) -> list[ToolCallItem] | None:
-    try:
-        text = text.removeprefix("<tool_call>")
-        text = text.removesuffix("</tool_call>")
-        top_level = {
-            k: json.dumps(v) if isinstance(v, (dict, list)) else v
-            for k, v in json.loads(text).items()  # pyright: ignore[reportAny]
-        }
-        return [ToolCallItem.model_validate(top_level)]
-    except Exception:
-        return None
-
-
-def _flatten(p: dict[str, Any]) -> dict[str, str]:
-    return {
-        k: json.dumps(v) if isinstance(v, (dict, list)) else str(v)  # pyright: ignore[reportAny]
-        for k, v in p.items()  # pyright: ignore[reportAny]
-    }
-
-
-json_tool_parser = ToolParser(
-    start_parsing="<tool_call>",
-    end_parsing="</tool_call>",
-    parse_tool_calls=_parse_json_calls,
-)
-
-
-def infer_tool_parser(chat_template: str) -> ToolParser | None:
-    """Attempt to auto-infer a tool parser from the chat template."""
-    if "<tool_call>" in chat_template and "tool_call.name" in chat_template:
-        return json_tool_parser
-    return None
--- a/src/exo/worker/tests/unittests/test_runner/test_event_ordering.py
+++ b/src/exo/worker/tests/unittests/test_runner/test_event_ordering.py
@@ -1,9 +1,7 @@
 # Check tasks are complete before runner is ever ready.
-import unittest.mock
 from collections.abc import Iterable
 from typing import Callable

-import mlx.core as mx
 import pytest

 import exo.worker.runner.runner as mlx_runner
@@ -21,7 +19,6 @@ from exo.shared.types.tasks import (
    Shutdown,
    StartWarmup,
    Task,
-    TaskId,
    TaskStatus,
    TextGeneration,
 )
@@ -116,7 +113,6 @@ def patch_out_mlx(monkeypatch: pytest.MonkeyPatch):
    monkeypatch.setattr(mlx_runner, "load_mlx_items", make_nothin((1, MockTokenizer)))
    monkeypatch.setattr(mlx_runner, "warmup_inference", make_nothin(1))
    monkeypatch.setattr(mlx_runner, "_check_for_debug_prompts", nothin)
-    monkeypatch.setattr(mlx_runner, "mx_any", make_nothin(False))
    # Mock apply_chat_template since we're using a fake tokenizer (integer 1).
    # Returns a prompt without thinking tag so detect_thinking_prompt_suffix returns None.
    monkeypatch.setattr(mlx_runner, "apply_chat_template", make_nothin("test prompt"))
@@ -167,7 +163,6 @@ def _run(tasks: Iterable[Task]):
    )

    task_sender, task_receiver = mp_channel[Task]()
-    _cancel_sender, cancel_receiver = mp_channel[TaskId]()
    event_sender = EventCollector()

    with task_sender:
@@ -178,16 +173,8 @@ def _run(tasks: Iterable[Task]):
        # this is some c++ nonsense
        task_receiver.close = nothin
        task_receiver.join = nothin
-        with unittest.mock.patch(
-            "exo.worker.runner.runner.mx.distributed.all_gather",
-            make_nothin(mx.array([1])),
-        ):
-            mlx_runner.main(
-                bound_instance,
-                event_sender,  # pyright: ignore[reportArgumentType]
-                task_receiver,
-                cancel_receiver,
-            )
+
+        mlx_runner.main(bound_instance, event_sender, task_receiver)  # type: ignore[arg-type]

        return event_sender.events

--- a/src/exo/worker/tests/unittests/test_runner/test_parse_tool_calls.py
+++ b/src/exo/worker/tests/unittests/test_runner/test_parse_tool_calls.py
@@ -5,13 +5,12 @@ from typing import Any

 from exo.shared.types.worker.runner_response import GenerationResponse, ToolCallResponse
 from exo.worker.runner.runner import parse_tool_calls
-from exo.worker.runner.tool_parsers import make_mlx_parser


 def _make_responses(
    texts: list[str],
    finish_on_last: bool = True,
-) -> Generator[GenerationResponse]:
+) -> Generator[GenerationResponse | ToolCallResponse]:
    """Create a sequence of GenerationResponses from text strings."""
    for i, text in enumerate(texts):
        is_last = i == len(texts) - 1
@@ -23,13 +22,10 @@ def _make_responses(
        )


-def _dummier_parser(text: str) -> dict[str, Any]:
+def _dummy_parser(text: str) -> dict[str, Any]:
    return {"name": "test_fn", "arguments": {"arg": text}}


-_dummy_parser = make_mlx_parser("<tool_call>", "</tool_call>", _dummier_parser)
-
-
 class TestParseToolCalls:
    """Tests for parse_tool_calls generator."""

@@ -39,6 +35,8 @@ class TestParseToolCalls:
        results = list(
            parse_tool_calls(
                _make_responses(texts, finish_on_last=False),
+                "<tool_call>",
+                "</tool_call>",
                _dummy_parser,
            )
        )
@@ -52,6 +50,8 @@ class TestParseToolCalls:
        results = list(
            parse_tool_calls(
                _make_responses(texts),
+                "<tool_call>",
+                "</tool_call>",
                _dummy_parser,
            )
        )
@@ -76,7 +76,9 @@ class TestParseToolCalls:
        results = list(
            parse_tool_calls(
                _make_responses(texts, finish_on_last=False),
-                make_mlx_parser("<tool_call>", "</tool_call>", _failing_parser),
+                "<tool_call>",
+                "</tool_call>",
+                _failing_parser,
            )
        )

--- a/tests/run_exo_on.sh
+++ b/tests/run_exo_on.sh
@@ -43,5 +43,4 @@ for host; do
  echo "Waiting for $host..."
  until curl -sf "http://$host:52415/models" &>/dev/null; do sleep 1; done
 done
-echo "all hosts alive!"
 wait
--- a/uv.lock
+++ b/uv.lock
@@ -377,8 +377,8 @@ dependencies = [
    { name = "hypercorn", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
    { name = "loguru", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
    { name = "mflux", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
-    { name = "mlx", version = "0.30.6", source = { registry = "https://pypi.org/simple" }, extra = ["cpu"], marker = "sys_platform == 'linux'" },
-    { name = "mlx", version = "0.30.7.dev20260217+50487b41", source = { git = "https://github.com/rltakashige/mlx-jaccl-fix-small-recv.git?branch=address-rdma-gpu-locks#50487b4141f3c951122655db3b83df5146c1fbeb" }, marker = "sys_platform == 'darwin'" },
+    { name = "mlx", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
+    { name = "mlx", extra = ["cpu"], marker = "sys_platform == 'linux'" },
    { name = "mlx-lm", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
    { name = "msgspec", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
    { name = "openai-harmony", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
@@ -416,9 +416,9 @@ requires-dist = [
    { name = "hypercorn", specifier = ">=0.18.0" },
    { name = "loguru", specifier = ">=0.7.3" },
    { name = "mflux", specifier = "==0.15.5" },
-    { name = "mlx", marker = "sys_platform == 'darwin'", git = "https://github.com/rltakashige/mlx-jaccl-fix-small-recv.git?branch=address-rdma-gpu-locks" },
+    { name = "mlx", marker = "sys_platform == 'darwin'", specifier = "==0.30.6" },
    { name = "mlx", extras = ["cpu"], marker = "sys_platform == 'linux'", specifier = "==0.30.6" },
-    { name = "mlx-lm", specifier = "==0.30.7" },
+    { name = "mlx-lm", specifier = "==0.30.6" },
    { name = "msgspec", specifier = ">=0.19.0" },
    { name = "openai-harmony", specifier = ">=0.0.8" },
    { name = "pillow", specifier = ">=11.0,<12.0" },
@@ -1020,8 +1020,8 @@ dependencies = [
    { name = "fonttools", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
    { name = "huggingface-hub", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
    { name = "matplotlib", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
-    { name = "mlx", version = "0.30.6", source = { registry = "https://pypi.org/simple" }, extra = ["cuda13"], marker = "sys_platform == 'linux'" },
-    { name = "mlx", version = "0.30.7.dev20260217+50487b41", source = { git = "https://github.com/rltakashige/mlx-jaccl-fix-small-recv.git?branch=address-rdma-gpu-locks#50487b4141f3c951122655db3b83df5146c1fbeb" }, marker = "sys_platform == 'darwin'" },
+    { name = "mlx", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
+    { name = "mlx", extra = ["cuda13"], marker = "sys_platform == 'linux'" },
    { name = "numpy", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
    { name = "opencv-python", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
    { name = "piexif", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
@@ -1048,12 +1048,18 @@ wheels = [
 name = "mlx"
 version = "0.30.6"
 source = { registry = "https://pypi.org/simple" }
-resolution-markers = [
-    "sys_platform == 'linux'",
+dependencies = [
+    { name = "mlx-metal", marker = "sys_platform == 'darwin'" },
 ]
 wheels = [
+    { url = "https://files.pythonhosted.org/packages/ae/5b/e460e144a34d5529e010056cccf50b538d56ed001473bc6b246018fd58cb/mlx-0.30.6-cp313-cp313-macosx_14_0_arm64.whl", hash = "sha256:ed86f8bffc174c2f259ca589ea25464c96cf69d1bb457074a2bf2ef53737e54f", size = 573515, upload-time = "2026-02-06T03:45:23.405Z" },
+    { url = "https://files.pythonhosted.org/packages/60/25/69833fefb9a3fef30b56792b1bcd022496c4fea83e45411d289b77ef7546/mlx-0.30.6-cp313-cp313-macosx_15_0_arm64.whl", hash = "sha256:c52294958269e20f300639a17c1900ca8fc737d859ddda737f9811e94bd040e5", size = 573516, upload-time = "2026-02-06T03:45:24.618Z" },
+    { url = "https://files.pythonhosted.org/packages/9c/6a/7e7fbeebc5cb51b6a5eba96b263a6298707bcbdc059f4b0b73e088bc3dea/mlx-0.30.6-cp313-cp313-macosx_26_0_arm64.whl", hash = "sha256:b5b6636f7c49a4d86d8ec82643b972f45a144a7a9f3a967b27b2e6e22cf71e6a", size = 573592, upload-time = "2026-02-06T03:45:25.928Z" },
    { url = "https://files.pythonhosted.org/packages/93/06/280f6f2ba80520a7109730425eda0d966658793aa0d02d8be8d351f75253/mlx-0.30.6-cp313-cp313-manylinux_2_35_aarch64.whl", hash = "sha256:67e6c9e30a9faeacc209917ef5523177cf9b086914b6b5d83ff886e4294b727d", size = 622011, upload-time = "2026-02-06T03:45:28.165Z" },
    { url = "https://files.pythonhosted.org/packages/fe/35/f872afbee9c079cc69924d9e9c46f5663adb7da58cba3511db082dd307c1/mlx-0.30.6-cp313-cp313-manylinux_2_35_x86_64.whl", hash = "sha256:47db8b16fcb6f6c5a47c0bdb24ed377b41237017ac93aa6cb6aa206c9bdf82e4", size = 663650, upload-time = "2026-02-06T03:45:30.315Z" },
+    { url = "https://files.pythonhosted.org/packages/60/23/361dc7a5797634e4d7e9bdd6564c6b28f9b1246672632def2f91bf066b18/mlx-0.30.6-cp314-cp314-macosx_14_0_arm64.whl", hash = "sha256:78804a89dcff4a838f7c2da72392fe87a523e95122a3c840e53df019122aad45", size = 575028, upload-time = "2026-02-06T03:45:31.549Z" },
+    { url = "https://files.pythonhosted.org/packages/a8/69/1854484d414171586814dfbe8def95f75c4ea2c7341ba13ba8ee675f7c62/mlx-0.30.6-cp314-cp314-macosx_15_0_arm64.whl", hash = "sha256:ec13584ab069665cc7ad34a05494d9291cd623aef6ae96be48875fc87cfc25d6", size = 575026, upload-time = "2026-02-06T03:45:33.072Z" },
+    { url = "https://files.pythonhosted.org/packages/6b/b8/3adbc441924209a7e4c568308b2a0b54bd09aee6a68db5bae85304791e54/mlx-0.30.6-cp314-cp314-macosx_26_0_arm64.whl", hash = "sha256:b2c5e8a090a753ef99a1380a4d059c983083f36198864f6df9faaf1223d083df", size = 575041, upload-time = "2026-02-06T03:45:34.814Z" },
    { url = "https://files.pythonhosted.org/packages/3f/54/9d9e06804fb2088202a2cdf60458e00b221f71420bea285720b60f9e82b5/mlx-0.30.6-cp314-cp314-manylinux_2_35_aarch64.whl", hash = "sha256:9ceddede4af0de31d1f6b3099f70e5469d60cd7c546975dedbdbeab3519cab3f", size = 624002, upload-time = "2026-02-06T03:45:36Z" },
    { url = "https://files.pythonhosted.org/packages/42/92/3140a15a50cb1f9267a6552171e1dfa577861de53e093124bc43707f2a0e/mlx-0.30.6-cp314-cp314-manylinux_2_35_x86_64.whl", hash = "sha256:4a6ffd2d16728cf95f63a1b555d7c2eaeea686a0e6b73228bd265411cb5d77a4", size = 663569, upload-time = "2026-02-06T03:45:37.242Z" },
 ]
@@ -1066,14 +1072,6 @@ cuda13 = [
    { name = "mlx-cuda-13", marker = "sys_platform == 'linux'" },
 ]

-[[package]]
-name = "mlx"
-version = "0.30.7.dev20260217+50487b41"
-source = { git = "https://github.com/rltakashige/mlx-jaccl-fix-small-recv.git?branch=address-rdma-gpu-locks#50487b4141f3c951122655db3b83df5146c1fbeb" }
-resolution-markers = [
-    "sys_platform == 'darwin'",
-]
-
 [[package]]
 name = "mlx-cpu"
 version = "0.30.6"
@@ -1100,20 +1098,30 @@ wheels = [

 [[package]]
 name = "mlx-lm"
-version = "0.30.7"
+version = "0.30.6"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
    { name = "jinja2", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
-    { name = "mlx", version = "0.30.7.dev20260217+50487b41", source = { git = "https://github.com/rltakashige/mlx-jaccl-fix-small-recv.git?branch=address-rdma-gpu-locks#50487b4141f3c951122655db3b83df5146c1fbeb" }, marker = "sys_platform == 'darwin'" },
+    { name = "mlx", marker = "sys_platform == 'darwin'" },
    { name = "numpy", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
    { name = "protobuf", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
    { name = "pyyaml", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
    { name = "sentencepiece", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
    { name = "transformers", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/66/0d/56542e2ae13ec6f542d3977d7cff89a205d4f6c5122e0ce23f33265f61c9/mlx_lm-0.30.7.tar.gz", hash = "sha256:e5f31ac58d9f2381f28e1ba639ff903e64f7cff1bdc245c0bc97f72264be329c", size = 275764, upload-time = "2026-02-12T18:41:11.86Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/76/cb/815deddc8699b1f694d7e1f9cbed52934c03a8b49432c8add72932bb2f0b/mlx_lm-0.30.6.tar.gz", hash = "sha256:807e042d7040268f1b19190b7eaefd8b2efbff5590a65460974ad4225b91dda1", size = 271733, upload-time = "2026-02-04T21:27:45.741Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/1e/17/a41c798a3d9cbdc47f39c6db5bba4c2cd199203ead26bf911cb03b644070/mlx_lm-0.30.7-py3-none-any.whl", hash = "sha256:17442a4bf01c4c2d3bca1e647712fe44f19890c3f1eadc8589d389e57b44b9bf", size = 386591, upload-time = "2026-02-12T18:41:10.236Z" },
+    { url = "https://files.pythonhosted.org/packages/20/5f/01d281f1fa8a1521d5936659beb4f5ab1f32b463d059263cf9d4cef969d9/mlx_lm-0.30.6-py3-none-any.whl", hash = "sha256:a7405bd581eacc4bf8209d7a6b7f23629585a0d7c6740c2a97e51fee35b3b0e1", size = 379451, upload-time = "2026-02-04T21:27:43.222Z" },
+]
+
+[[package]]
+name = "mlx-metal"
+version = "0.30.6"
+source = { registry = "https://pypi.org/simple" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/f3/85/44406b521f920248fad621334d4dc15e77660a494edf890e7cbee33bf38d/mlx_metal-0.30.6-py3-none-macosx_14_0_arm64.whl", hash = "sha256:ea6d0c973def9a5b4f652cc77036237db3f88c9d0af63701d76b5fddde99b820", size = 38437818, upload-time = "2026-02-06T03:44:56.19Z" },
+    { url = "https://files.pythonhosted.org/packages/d0/cb/10a516995f7d0c154b0d7e633c54b51e96977a86a355105b6474cfcbe0d0/mlx_metal-0.30.6-py3-none-macosx_15_0_arm64.whl", hash = "sha256:0f8cb94634d07e06a372d6ad9a090f38a18bab1ff19a140aede60eacf707bb94", size = 38433701, upload-time = "2026-02-06T03:44:59.678Z" },
+    { url = "https://files.pythonhosted.org/packages/4c/7d/70cb272f7373c334709f210ed8420511fc9d64d05a7a646c0b3b94c29c04/mlx_metal-0.30.6-py3-none-macosx_26_0_arm64.whl", hash = "sha256:d761ae26304f2c4b454eeea7f612a56919d9e5e57dbb1dc0788f8e34aa6f41c2", size = 47718448, upload-time = "2026-02-06T03:45:03.133Z" },
 ]

 [[package]]
Author	SHA1	Message	Date
Alex Cheema	03f5dafa2a	Merge remote-tracking branch 'origin/main' into alexcheema/prefill-progress	2026-02-13 09:37:55 -08:00
Alex Cheema	92401ab7f8	chore: fix formatting after merge Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>	2026-02-13 06:54:18 -08:00
Alex Cheema	d0c8501cb0	Merge remote-tracking branch 'origin/main' into alexcheema/prefill-progress # Conflicts: # src/exo/master/adapters/chat_completions.py # src/exo/worker/engines/mlx/generator/generate.py # src/exo/worker/runner/runner.py	2026-02-13 05:58:45 -08:00
Alex Cheema	f94870f2c0	Merge remote-tracking branch 'origin/main' into alexcheema/prefill-progress # Conflicts: # dashboard/src/lib/stores/app.svelte.ts	2026-02-05 05:38:14 -08:00
Alex Cheema	18928d4d9e	Merge remote-tracking branch 'origin/main' into alexcheema/prefill-progress	2026-02-05 05:30:22 -08:00
Alex Cheema	937da476b0	address PR review comments: use model_dump_json, match/case, prefill_step_size=1024 - Use model_dump_json() for prefill progress SSE instead of hand-crafted JSON - Convert isinstance chain to match/case in generate_chat_stream - Set all prefill_step_size to 1024 (was 256 for testing, 2048 default) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-02-03 09:04:25 -08:00
Alex Cheema	1c1c286127	Merge uncertainty-visualization into prefill-progress Combines prefill progress bar (PrefillProgressData, SSE named events) with uncertainty visualization features (logprobs, tool calls, error chunks, traces, image generation). Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-02-03 06:18:54 -08:00
Alex Cheema	258785be84	Merge remote-tracking branch 'origin/main' into alexcheema/uncertainty-visualization	2026-02-03 06:03:01 -08:00
Alex Cheema	13a6b9819a	fix: assistant prefilling for regenerate-from-token and tooltip UX Support assistant message continuation by popping the last assistant message before template formatting and appending its content raw, keeping the turn open without a closing token. Improve tooltip hover UX: use getClientRects() for correct multi-line token positioning, add padding to bridge the hover gap, and increase the hide delay. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-02-03 06:00:28 -08:00
Alex Cheema	1733d07cb3	fix: enable uncertainty visualization for regular chat messages The sendMessage method was missing logprobs request params and token collection, so the heatmap toggle never appeared. Also rename the top_k parameter to top_logprobs in extract_top_logprobs to avoid confusion with the sampling top_k parameter. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-02-03 05:08:49 -08:00
Alex Cheema	b3e4c9b1e5	fix: populate logprobs in non-streaming chat completions responses collect_chat_response() was dropping logprobs data from TokenChunks, so non-streaming requests never returned logprobs even when requested. Accumulate LogprobsContentItems and attach them to the ChatCompletionChoice. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-02-03 04:45:39 -08:00
Alex Cheema	4c74792373	Merge branch 'main' into alexcheema/uncertainty-visualization	2026-02-03 04:44:14 -08:00
Alex Cheema	eadb6de1f7	Merge main into uncertainty-visualization branch Resolve conflicts by keeping main's structure (TextGenerationTaskParams, tool calling, KV prefix cache, Claude/OpenAI APIs) and surgically adding the uncertainty visualization features on top: - Add logprob/top_logprobs fields to GenerationResponse and TokenChunk - Add extract_top_logprobs() to MLX generator for per-token logprob extraction - Build Logprobs in chat completions adapter for streaming responses - Add SSE headers (Cache-Control, Connection, X-Accel-Buffering) to streaming endpoints - Add TokenHeatmap component and uncertainty toggle in dashboard - Add logprobs collection in streaming response handler - Add regenerateFromToken method for re-generation from specific tokens - Strip token data from localStorage to avoid storage bloat Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-02-02 11:33:24 -08:00
Alex Cheema	7ba2408eed	fix: restore dashboard build by using main's app.svelte.ts The prefill-progress branch's app.svelte.ts was missing image generation features from main. To fix the dashboard build, restored main's app.svelte.ts and removed the uncertainty visualization and prefill progress bar features from ChatMessages.svelte that depended on the missing exports. Note: TokenHeatmap and PrefillProgressBar components still exist but are not currently used. The prefill progress backend code is still in place and can be re-enabled in the dashboard once app.svelte.ts is properly updated to include both image generation and prefill/uncertainty features. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 12:07:16 +00:00
Alex Cheema	ce4d7f4d43	style: simplify prefill progress bar and use exo color palette - Remove spinner (progress bar is dynamic enough) - Use exo-yellow for progress bar fill - Use exo-black/60 for progress bar background - Use exo-light-gray for text Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 12:00:55 +00:00
Alex Cheema	6727523eab	fix: wire prefill progress events to chat completions stream - Move PrefillProgressData to shared types (chunks.py) to avoid circular imports - Update generate_chat_stream adapter to handle both TokenChunk and PrefillProgressData - Use _stream_events instead of _chat_chunk_stream for streaming endpoint - Prefill progress now properly sent as SSE 'event: prefill_progress' to frontend Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 12:00:55 +00:00
Alex Cheema	a8f81e0495	feat: add prefill progress bar for long prompts Shows real-time progress during prompt processing (prefill phase). Progress is sent via SSE named events that maintain OpenAI API compatibility. - Add PrefillProgress event type and PrefillProgressData dataclass - Wire prompt_progress_callback through MLX stream_generate - Send progress events directly from callback for real-time updates - Add PrefillProgressBar.svelte component - Parse event: prefill_progress SSE events in dashboard Note: prefill_step_size temporarily set to 256 for testing (normally 2048) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:59:52 +00:00
Alex Cheema	ba7148ccec	style: format app.svelte.ts with nix fmt	2026-01-22 11:53:43 +00:00
Alex Cheema	a64b8addc6	Fix localStorage quota issues by stripping tokens and auto-pruning - Strip tokens (logprobs data) from messages before saving to localStorage since they're large and not essential for persistence - Add pruneOldConversations() to automatically remove oldest conversations when quota is exceeded - This prevents QuotaExceededError from crashing the app Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:53:43 +00:00
Alex Cheema	e6599a9408	Fix ReferenceError: controller undefined in sendMessage finally block Move AbortController creation before the try block in both sendMessageWithLogprobs and regenerateFromToken functions. Previously, controller was defined inside the try block but referenced in the finally block, causing a ReferenceError if an exception was thrown before the controller was created. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:53:43 +00:00
Alex Cheema	93f4753598	Add SSE headers to properly close streaming connections Add Cache-Control, Connection: close, and X-Accel-Buffering headers to all SSE streaming responses. These headers help ensure: - No caching of streaming responses - Connection closes when stream ends (instead of keep-alive) - No proxy buffering that could delay stream closure This should fix the issue where the frontend stays on "PROCESSING" even after receiving the complete response. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:53:43 +00:00
Alex Cheema	75fe505275	Add debug logging to generate_chat_stream Add logging to help diagnose why streaming might not be ending properly. This will show when [DONE] is yielded, when return is called, and when the finally block runs. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:53:43 +00:00
Alex Cheema	d7c044e349	Fix streaming not ending after [DONE] is yielded Add missing return statement after yielding [DONE] in generate_chat_stream. Without this, the async generator continues waiting for more chunks from chunk_stream even though generation is complete, causing the stream to hang indefinitely. The frontend waits for the stream to close (reader.done) which never happens, resulting in the chat button staying on "PROCESSING" forever. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:53:43 +00:00
Alex Cheema	53b6d56e9f	fix: restore extract_top_logprobs function for uncertainty visualization The extract_top_logprobs function was lost during rebases. This function processes the out.logprobs array (full vocabulary logprobs from MLX) to extract the selected token's logprob and top-k alternatives. The previous code tried to use getattr(out, "logprob", None) which doesn't exist - mlx_lm returns logprobs as an mx.array, not individual values. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:53:43 +00:00
Alex Cheema	7fe0a61230	fix: remove unsupported logprob params from stream_generate The mlx_lm.stream_generate already returns logprobs in its output - we don't need to pass return_logprob or return_top_logprobs kwargs. The uncertainty visualization feature extracts logprobs from the existing out.logprobs field. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:53:43 +00:00
Alex Cheema	5a36542631	feat: add uncertainty visualization with token-level logprobs - Add TokenHeatmap component for visualizing token confidence - Collect and stream logprobs in generation pipeline - Add regenerate-from-token feature with continue_from_prefix - Add AbortController for request cancellation - Support continue_final_message for seamless prefix continuation Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:53:43 +00:00
Alex Cheema	955e0105b3	fix: resolve import and type errors from rebase - Use claude_request_to_internal instead of old function name - Fix ModelId imports in runner.py and test files - Update test_mlx/conftest.py to use ResponsesRequest format - Remove unused imports Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:36:11 +00:00
Evan	4d1eb1d9bd	fix: rebase fix	2026-01-22 11:32:46 +00:00
Alex Cheema	365416c65e	style: move inline imports to top of file in api.py Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:32:26 +00:00
Alex Cheema	04af76e10f	fix: restore try/except structure in runner.py Replace non-existent context manager with proper try/except block and remove unused ModelId import. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:32:04 +00:00
Alex Cheema	a84c3431cd	style: fix formatting issues caught by treefmt Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:31:45 +00:00
Alex Cheema	52445b21f6	refactor: use ResponsesRequest as canonical internal type - Extend ResponsesRequest with fields: top_k, seed, stop, tools - Remove redundant InternalTaskParams and InputMessage types - Update all adapters to convert to ResponsesRequest - Simplify Responses API (no conversion needed - native passthrough) - Update all imports across codebase and tests This eliminates type duplication and makes the Responses API relationship explicit throughout the codebase. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:31:44 +00:00
Alex Cheema	435bd7f6fa	refactor: make Responses API the canonical internal format Restructure the API layer so that OpenAI Responses API is the native format, with Chat Completions and Claude Messages as adapters on top. Changes: - Add new chat_completions.py adapter with streaming/non-streaming support - Update responses.py with collect_responses_response() for non-streaming - Update claude.py with collect_claude_response() for non-streaming - Refactor api.py so all endpoints use adapters uniformly - Rename _chat_chunk_stream to _token_chunk_stream (generic internal format) - Remove unused chat_response_to_* converter functions - Update tests to remove tests for deleted functions Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:30:27 +00:00
Alex Cheema	dd25b5b90e	feat: add Claude Messages API and OpenAI Responses API support Adds two new API endpoints that wrap the existing chat completions: - /v1/messages - Claude Messages API compatible endpoint - /v1/responses - OpenAI Responses API compatible endpoint Both support streaming (SSE) and non-streaming modes with proper token usage reporting from actual inference stats. Also adds top_k sampling parameter and stop sequence support to the MLX inference engine. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-01-22 11:28:49 +00:00