mirror of
https://github.com/mudler/LocalAI.git
synced 2026-07-30 09:57:57 -04:00
49ef40a187
* feat(llama-cpp): route Score through the slot loop Score previously bypassed the slot loop with a direct llama_decode: a conflict guard aborted the whole process if scoring raced generation, the config validator had to reject score alongside chat/completion/embeddings, and every candidate re-decoded the full shared prompt. Add SERVER_TASK_TYPE_SCORE to the (patched) upstream server so score tasks are scheduled like any other slot work: generation and scoring serialize naturally, the shared prompt is decoded once per call, and the slot's prompt cache carries the conversation prefix across calls. Context checkpoints at the score boundary and at the cache-divergence point keep SWA/hybrid/recurrent models (e.g. LFM2.5) from re-prefilling the whole prompt per candidate: warm-turn scoring on a 6-option set drops from ~8s to ~0.5s on a desktop CPU. The conflict guard and the validation split are removed; declaring score with generation usecases on one config is now supported and shares the slot cache. Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * feat(realtime): classifier wire types and pipeline config Wire types and YAML config for realtime classifier mode: sessions carry a localai_classifier extension (options with canned replies/tool calls, softmax threshold, normalization, history trimming, fallback modes, and a deterministic wake-word address gate), mirrored by pipeline.classifier in the model YAML and surfaced in the config-meta registry. The localai.classifier.result server event reports the full score distribution per turn. Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * feat(realtime): classifier response flow Classifier-mode responses: instead of autoregressive generation, each user turn is prefill-scored against the option list (router.ScoreClassifier prompt/candidate shapes over the Score primitive) and the winning option's canned reply and tool call are emitted through the existing response machinery. Below-threshold turns take the configured fallback (none / canned reply / generate); empty transcripts and unaddressed turns (wake word not mentioned) skip scoring entirely. The scoring probe defaults to the latest user message only — small scorers echo canned replies from prior turns back as the top option otherwise. Built for hardware that can afford prompt processing but not decode: with slot-based Score the option list stays KV-cached across turns, so a turn costs roughly one forward pass over the new words. session_update_error events now carry the validation cause instead of a generic message. Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * fix(realtime): bound the VAD tick's scan window and buffer retention The VAD tick loop re-scanned the entire input buffer every 300ms and only trimmed it on zero-segment ticks or commits. Audio that keeps producing segments without a committing pause (steady noise a mic pipeline lets through, music, continuous speech) grew the buffer toward the 100MB cap with each tick rescanning all of it — O(n^2), measured at ~3.3ms of silero per buffered second: past ~90s retained, ticks run back to back and pin ~4 cores until the stream stops. Silero's recurrent state only carries a few hundred ms of context, so rescanning old audio buys nothing. Clip the slice handed to the VAD to the largest silence the commit test can need to measure (server_vad silence window or the semantic eagerness fallback) plus a warm-up margin, and rebase the returned segment times so every downstream consumer keeps whole-buffer coordinates. An open turn whose clipped window is all silence now commits (the silence outran the window) instead of being discarded as no-speech. Independently, retain at most 90s of raw buffer, rebasing the live-feed and EOU cursors on trim — this also bounds the previously unbounded VAD-error path. Turn boundaries are otherwise unchanged: no forced commits, no new coordinator states. pipeline.turn_detection.vad_window_sec can widen the scan window; values below the automatic floor are ignored. The tick body is extracted into vadTick so specs can drive turn detection synchronously (same shape as classifySoundWindow); the babble reproduction that pinned 4 cores now plateaus under 10% of one core. Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * fix(backend): let per-model threads override the global default ModelOptions overrode a set per-model threads value with the app-level --threads whenever the latter was non-zero — and WithThreads defaults it to the physical core count, so it always was. The YAML threads: knob has been dead config: a tiny VAD model could never opt down from the global pool size. SetDefaults already fills an unset per-model value from the app config, which is the intended precedence; resolve threads through a helper that honors it (explicit threads: 0 still means unset). Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * chore(gallery): single-thread the silero VAD Silero is a ~2MB recurrent model with no exploitable graph parallelism: measured per-call latency is identical at 1 and 10 ORT threads, while every extra pool thread just spin-waits between the realtime loop's frequent tiny inferences. Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * docs(realtime): classifier mode, VAD scan window, threads precedence Document the realtime classifier mode (options, threshold guidance, wake-word address gate, empty-transcript handling), the VAD scan window and 90s buffer retention (pipeline.turn_detection.vad_window_sec), the per-model threads precedence, and the M3 classifier note in the realtime state-machine design doc. Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * perf(llama-cpp): score all candidates in one batched decode One scoring call is now a single SERVER_TASK_TYPE_SCORE task: the slot decodes the shared prefix (prompt + longest common candidate token prefix) once, then forks one sequence per candidate off it (metadata-only for the unified KV cache, copy-on-write for recurrent state) and decodes every candidate's unique tail in one llama_decode. Previously each candidate was its own task that restored the boundary checkpoint and re-decoded its full tail sequentially, paying per-candidate task and decode overhead. The context reserves SERVER_SCORE_FORK_SEQS extra sequence ids (and recurrent-state cells) beyond the parallel slots via the new common_params::n_seq_score_forks. Forking requires the unified KV cache (already this backend's default) since per-sequence streams would shrink n_ctx_seq; an explicit kv_unified:false disables forking and Score calls that need it fail cleanly. Candidates beyond the fork/output budget decode in successive chunks. Wire contract and scores are unchanged: per-token logprobs are stitched from the shared region and the forked tails. Verified bitwise deterministic call-to-call and independent of candidate order (no cross-fork leakage via equal-length candidate swap); ranking matches the per-candidate implementation on the drone battery (winner softmax 0.99996 vs 0.99997), and >16-candidate chunking, prefix-of-another and empty candidates all pass. Measured on a desktop CPU: warm /api/score calls 0.52s -> 0.23s; warm realtime classifier turns 196-303ms. The 9-candidate drone turn decodes ~17 unique tail tokens in one batch instead of nine sequential ~220ms checkpoint-restore tasks. Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * fix(realtime): gate scoring capacity by model usecase Reserve llama.cpp scoring slots only for models that explicitly declare the score usecase, while allowing score to coexist with chat and completion. Reject incompatible unified-KV settings and classifier activation on models without scoring capacity. Propagate application defaults when resolving realtime and preload pipeline stages so unset thread counts are resolved consistently without overriding explicit model settings. Assisted-by: Codex:gpt-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * fix(ci): honor APT mirrors in the prebuilt llama-cpp compile step The builder-prebuilt path installs gcc-14 with apt directly and ignored the APT_MIRROR/APT_PORTS_MIRROR build args the from-source path already honors, so an ubuntu mirror outage broke every arm64 backend build. Pass the args into the stage and run apt-mirror.sh (already in the build context via COPY . /LocalAI) before the apt step. Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * feat(realtime): classifier argument slots via constrained completion Hybrid classify-then-complete: a classifier option's canned tool call can declare typed argument slots (number | enum | string, with defaults and prompt hints) referenced as "{{name}}" in the arguments template. When the option wins, the slots are filled by a short grammar-constrained completion that continues the exact scoring prompt — rendered by the same cached ScoreClassifier, so the llama.cpp prompt cache is already warm — with the chosen route JSON re-opened at the first slot field. A GBNF grammar pins the field skeleton and frees only the values; temperature 0, a couple dozen tokens at most (~300ms on a desktop CPU for two slots). Slot declarations and hints ride the option descriptions in the shared system prompt, informing scoring and the fill alike at no per-turn token cost. The localai.classifier.result event carries the final arguments and a fill_latency_ms. On inference failure the slots' defaults apply; a slot without a default fails the response (or falls through with fallback.mode: generate). Slot filling requires completion alongside score in the scoring model's known_usecases. Verified end-to-end on the Pi drone demo: "fly forward three meters" in distance mode classifies forward and infers {"distance": 3, "units": "meters"} in ~310ms, and the drone flies exactly 3 units. Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * feat(realtime): splice filled slot values into classifier replies A classifier option's spoken reply can now reference its tool's argument slots ("Going forward {{distance}} {{units}}."): the values inferred by the slot-fill completion — or the recovery defaults — are spliced into the reply as plain text before it is emitted, so what the assistant says confirms what it actually inferred. Placeholders without a value stay literal, and options without slots are untouched. FillToolArguments now returns the raw slot values alongside the spliced arguments JSON to make the reply templating possible. Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * fix(realtime): harden classifier slot completion Reserve context for constrained slot filling, size completions from their encoded output, and encode enum grammar literals as valid JSON. Reject empty enum values and cover the failure modes with regression tests. Assisted-by: Codex:gpt-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * feat(realtime): prewarm the classifier scoring prompt on registration Swapping a session's classifier option list (a voice-switched command mode, for instance) made the next turns pay a full re-prefill of the new option-list prompt — measured 2.4s vs 0.3s warm on a desktop CPU, and worse: on hybrid-memory models like LFM2.5, whose state cannot be partially rewound (llama.cpp can only restore checkpoints), *every* probe change re-prefilled from scratch whenever the last checkpoint missed the probe boundary, so even same-list turns intermittently cost full prefills. Registering an option list (pipeline seed or session.update) now fires a best-effort background prewarm: two throwaway scores with distinct probes. The first prefills the new option-list prompt; the second, diverging exactly where per-turn probe text starts, plants the backend's rewind point (KV checkpoint) at the stable-prefix boundary that every real turn reuses. The prewarm hides behind the canned mode-switch reply — by the time it finishes speaking, the cache is warm. Idempotent per option set, detached from the registering request's lifetime. Measured on the drone demo (LFM2.5-1.2B, desktop CPU): first turn after a mode switch 2374ms -> 340ms; intermittent same-list full prefills (1.3-2.1s) all -> under 0.5s. For clients that swap lists frequently, options: [parallel:2] on the scoring model additionally keeps one slot per list via prefix-similarity routing (+26MB RSS, unified KV). Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * perf(llama-cpp): checkpoint scoring at the caller-declared stable prefix Hybrid-memory models (LFM2.5 shortconv, Qwen3.5 deltanet — where new small models are headed) cannot rewind their state, so any prompt-cache reuse that needs a rewind falls back to a full re-prefill. For classifier scoring that meant every probe change re-processed the whole option-list prompt: the server's checkpoints were placed reactively (at wherever the previous task happened to diverge), so a checkpoint past the next divergence was erased rather than restored — measured as intermittent 2-10s turns on prompts with a 95%+ common prefix. The classifier now computes the probe-invariant prompt prefix once (the byte-wise common prefix of two synthetic probe renders) and declares its length with every Score request; the server maps it to a token boundary and forces a KV checkpoint exactly there on each score prefill. That checkpoint sits at or before every future divergence under the same option list, so it always survives and always restores — repeat scoring costs probe+candidates regardless of how the probe changes. Also: - prewarm reruns on every option-list registration instead of memoizing per list: with boundary checkpoints a redundant rewarm costs two probe-sized decodes, while skipping one after a slot eviction (three lists sharing fewer slots evict in LRU cascades) silently moves a full re-prefill onto the user's next turn - new llama.cpp backend option rs_seq:N exposes bounded recurrent-state rollback outside speculative decoding; measured impractical for deltanet-scale states (65GB for 64 snapshots on Qwen3.5-4B) but cheap insurance for small-state models - docs: the multi-list recipe (parallel:N + sps:0.5 — the default slot similarity threshold funnels distinct lists onto one slot) Measured on the drone demo (LFM2.5-1.2B scorer, desktop CPU), steady state: every turn 285-421ms including mode switches, vs 2.4s post-switch and intermittent 1.3-2.9s re-prefills before. Assisted-by: Claude:claude-fable-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * fix(realtime): align classifier cache guidance Document the single-score prewarm behavior and clean the vendored score patch formatting. Assisted-by: Codex:gpt-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> * fix(llama-cpp): guard score task for fork backends TurboQuant and Bonsai reuse the primary gRPC server against llama.cpp forks that do not carry LocalAI's slot-based Score patches. Compile the Score integration only for the patched primary backend and return UNIMPLEMENTED from fork builds instead of referencing absent task types and common_params fields. Assisted-by: Codex:gpt-5 [gh] Signed-off-by: Richard Palethorpe <io@richiejp.com> * fix(dev): generate gRPC code before commit lint The coverage phase regenerates ignored protobuf bindings, but lint runs first and can fail against missing or stale output. Generate the pinned bindings before lint so the gate always type-checks the current schema. Assisted-by: Codex:gpt-5 Signed-off-by: Richard Palethorpe <io@richiejp.com> --------- Signed-off-by: Richard Palethorpe <io@richiejp.com>
Deutsch | Español | français | 日本語 | 한국어 | Português | Русский | 中文
LocalAI is the open-source AI engine. Run any model - LLMs, vision, voice, image, video - on any hardware. No GPU required.
A small core, not a bundle. Each backend wraps a best-in-class engine (llama.cpp, vLLM, whisper.cpp, stable-diffusion, MLX...) in its own image, pulled only when a model needs it. You install nothing you don't use.
- Composable by design: backends are separate and pulled on demand, so you install only what your model needs
- Open and extensible: load any model, or build your own backend in any language against an open interface
- Drop-in API compatibility: OpenAI, Anthropic, and ElevenLabs APIs across every backend
- Any model, any modality: LLMs, vision, voice, image, and video behind one API
- Any hardware: NVIDIA, AMD, Intel, Apple Silicon, Vulkan, or CPU-only
- Multi-user ready: API key auth, user quotas, role-based access
- Built-in AI agents: autonomous agents with tool use, RAG, MCP, and skills
- Privacy-first: your data never leaves your infrastructure
Created by Ettore Di Giacinto and maintained by the LocalAI team.
📖 Documentation | 💬 Discord | 💻 Quickstart | 🖼️ Models | ❓FAQ
Guided tour
https://github.com/user-attachments/assets/08cbb692-57da-48f7-963d-2e7b43883c18
Click to see more!
User and auth
https://github.com/user-attachments/assets/228fa9ad-81a3-4d43-bfb9-31557e14a36c
Agents
https://github.com/user-attachments/assets/6270b331-e21d-4087-a540-6290006b381a
Usage metrics per user
https://github.com/user-attachments/assets/cbb03379-23b4-4e3d-bd26-d152f057007f
Fine-tuning and Quantization
https://github.com/user-attachments/assets/5ba4ace9-d3df-4795-b7d4-b0b404ea71ee
WebRTC
https://github.com/user-attachments/assets/ed88e34c-fed3-4b83-8a67-4716a9feeb7b
Quickstart
macOS
Note: The DMG is not signed by Apple. After installing, run:
sudo xattr -d com.apple.quarantine /Applications/LocalAI.app. See #6268 for details.
Containers (Docker, podman, ...)
Already ran LocalAI before? Use
docker start -i local-aito restart an existing container.
CPU only:
docker run -ti --name local-ai -p 8080:8080 localai/localai:latest
NVIDIA GPU:
# CUDA 13
docker run -ti --name local-ai -p 8080:8080 --gpus all localai/localai:latest-gpu-nvidia-cuda-13
# CUDA 12
docker run -ti --name local-ai -p 8080:8080 --gpus all localai/localai:latest-gpu-nvidia-cuda-12
# NVIDIA Jetson ARM64 (CUDA 12, for AGX Orin and similar)
docker run -ti --name local-ai -p 8080:8080 --gpus all localai/localai:latest-nvidia-l4t-arm64
# NVIDIA Jetson ARM64 (CUDA 13, for DGX Spark)
docker run -ti --name local-ai -p 8080:8080 --gpus all localai/localai:latest-nvidia-l4t-arm64-cuda-13
AMD GPU (ROCm):
docker run -ti --name local-ai -p 8080:8080 --device=/dev/kfd --device=/dev/dri --group-add=video localai/localai:latest-gpu-hipblas
Intel GPU (oneAPI):
docker run -ti --name local-ai -p 8080:8080 --device=/dev/dri/card1 --device=/dev/dri/renderD128 localai/localai:latest-gpu-intel
Vulkan GPU:
docker run -ti --name local-ai -p 8080:8080 localai/localai:latest-gpu-vulkan
Loading models
# From the model gallery (see available models with `local-ai models list` or at https://models.localai.io)
local-ai run llama-3.2-1b-instruct:q4_k_m
# From Huggingface
local-ai run huggingface://TheBloke/phi-2-GGUF/phi-2.Q8_0.gguf
# From the Ollama OCI registry
local-ai run ollama://gemma:2b
# From a YAML config
local-ai run https://gist.githubusercontent.com/.../phi-2.yaml
# From a standard OCI registry (e.g., Docker Hub)
local-ai run oci://localai/phi-2:latest
To test a running LocalAI server from the terminal, open an interactive chat session from another shell. Inside the prompt, /models lists installed models and /model <name> switches between them.
# Terminal 1
local-ai run llama-3.2-1b-instruct:q4_k_m
# Terminal 2
local-ai chat --model llama-3.2-1b-instruct:q4_k_m
Automatic Backend Detection: LocalAI automatically detects your GPU capabilities and downloads the appropriate backend. For advanced options, see GPU Acceleration.
For more details, see the Getting Started guide.
Latest News
- June 2026: New native biometric backends from the LocalAI team: voice-detect.cpp for speaker recognition and voice analysis (ECAPA-TDNN, WeSpeaker, ERes2Net, CAM++, wav2vec2 age/gender/emotion) and face-detect.cpp for face detection, recognition, demographics and anti-spoofing (SCRFD/ArcFace, YuNet/SFace). Both are from-scratch C++/ggml engines with no Python or onnxruntime at inference, self-contained GGUF weights, bit-exact parity with the reference, and GPU cuDNN parity, replacing the heavier Python
insightfaceandspeaker-recognitionbackends (PR #10441). - June 2026: New realtime voice assistant demo (a tiny Go client for the Realtime API with a full talk-back voice loop and tool calling), plus streaming of the realtime LLM / TTS / transcription pipeline stages and configurable WebRTC ICE candidates.
- June 2026: Big speech push: the parakeet.cpp ASR engine gains NeMo-faithful segment timestamps, a multilingual streaming Nemotron-3.5 model, dynamic batching for concurrent transcription and CUDA graphs; the new CrispASR backend adds multi-architecture ASR + TTS, and 60 Piper TTS voices across 42 languages land in the gallery (plus per-request TTS instructions and params).
- June 2026: New backends and models: locate-anything.cpp for open-vocabulary object detection via ggml, Ideogram4 image generation in stablediffusion-ggml, llama.cpp video input, and the Gemma 4 QAT family with MTP speculative-decoding pairs. Plus an interactive CLI chat mode and RAG source citations in agent responses.
- June 2026: Distributed mode hardening: prefix-cache-aware routing, a production-ready request router with auto-sized embedding/rerank batches, ds4 layer-split distributed inference, NATS JWT auth + TLS/mTLS, and resumable file uploads.
- May 2026: LocalAI 4.3.0 -
llama.cppprompt cache on by default (repeated system prompts collapse from minutes to seconds), keyless cosign signing of backend OCI images, per-API-key + per-user usage attribution, Distributed v3 with per-request replica routing. Release notes - May 2026: LocalAI 4.2.0 - LocalAI sees and hears: voice recognition, face recognition + antispoofing liveness, speaker diarization. Plus drop-in Ollama API, video generation, redesigned UI with i18n + admin-configurable branding, vLLM at feature parity with llama.cpp, and 11 new backends. Release notes
- April 2026: LocalAI 4.1.0 - LocalAI becomes a control tower: distributed cluster mode with VRAM-aware smart routing + autoscaling, multi-user platform with OIDC and API keys, per-user quotas with predictive analytics, in-UI fine-tuning with TRL (auto-export to GGUF), on-the-fly quantization backend, visual pipeline editor. Release notes
- March 2026: LocalAI 4.0.0 - native agentic orchestration with the new Agenthub community hub, full React UI rewrite with Canvas mode, MCP Apps + client-side with tool streaming, WebRTC realtime audio, MLX-distributed. Release notes
- February 2026: Realtime API for audio-to-audio with tool calling, ACE-Step 1.5 support
- January 2026: LocalAI 3.10.0 — Anthropic API support, Open Responses API, video & image generation (LTX-2), unified GPU backends, tool streaming, Moonshine, Pocket-TTS. Release notes
- December 2025: Dynamic Memory Resource reclaimer, Automatic multi-GPU model fitting (llama.cpp), Vibevoice backend
- November 2025: Import models via URL, Multiple chats and history
- October 2025: Model Context Protocol (MCP) support for agentic capabilities
- September 2025: New Launcher for macOS and Linux, extended backend support for Mac and Nvidia L4T, MLX-Audio, WAN 2.2
- August 2025: MLX, MLX-VLM, Diffusers, llama.cpp now supported on Apple Silicon
- July 2025: All backends migrated outside the main binary — lightweight, modular architecture
For older news and full release notes, see GitHub Releases and the News page.
Features
- Text generation (
llama.cpp,transformers,vllm... and more) - Text to Audio
- Audio to Text
- Image generation
- OpenAI-compatible tools API
- Realtime API (Speech-to-speech)
- Embeddings generation
- Constrained grammars
- Download models from Huggingface
- Vision API
- Object Detection
- Reranker API
- P2P Inferencing
- Distributed Mode — Horizontal scaling with PostgreSQL + NATS
- Model Context Protocol (MCP)
- Built-in Agents — Autonomous AI agents with tool use, RAG, skills, SSE streaming, and Agent Hub
- Backend Gallery — Install/remove backends on the fly via OCI images
- Voice Activity Detection (Silero-VAD)
- Integrated WebUI
Supported Backends & Acceleration
LocalAI supports 60+ backends including llama.cpp, vLLM, SGLang, transformers, whisper.cpp, diffusers, MLX, MLX-VLM, and many more. Hardware acceleration is available for NVIDIA (CUDA 12/13), AMD (ROCm), Intel (oneAPI/SYCL), Apple Silicon (Metal), Vulkan, and NVIDIA Jetson (L4T). All backends can be installed on-the-fly from the Backend Gallery.
See the full Backend & Model Compatibility Table and GPU Acceleration guide.
Backends built by us
Most backends wrap a best-in-class upstream engine. A handful of them are native C/C++/GGML engines (no Python at inference) developed and maintained by the LocalAI project itself:
| Backend | What it does |
|---|---|
| vllm.cpp | From-scratch C++20 port of vLLM for text generation: paged KV cache, continuous batching, prefix caching, safetensors + GGUF loading, engine-enforced structured output, on CPU, CUDA, Metal and Vulkan |
| parakeet.cpp | C++/GGML port of NVIDIA NeMo Parakeet ASR (tdt/ctc/rnnt/hybrid), with cache-aware streaming transcription |
| moss-transcribe.cpp | C++/GGML port of OpenMOSS MOSS-Transcribe-Diarize: joint long-form transcription, speaker diarization and timestamping in a single pass |
| moss-tts.cpp | C++/GGML port of the OpenMOSS MOSS-TTS family: text-to-speech (MOSS-TTS-Local v1.5, 48 kHz stereo) with reference-audio voice cloning, through the MOSS-Audio-Tokenizer neural codec |
| magpie-tts.cpp | C++/GGML port of NVIDIA's Magpie TTS Multilingual 357M: 22.05 kHz mono text-to-speech in 5 voices and 9+ languages, with the NanoCodec neural codec and tokenizer/G2P embedded in a single GGUF |
| ced.cpp | C++/GGML port of the CED audio-tagging models: sound-event classification (527-class AudioSet) over REST and the realtime API for live recognition |
| voice-detect.cpp | Speaker recognition and voice analysis (ECAPA-TDNN, WeSpeaker, ERes2Net, CAM++, wav2vec2 age/gender/emotion), replacing the Python speaker-recognition backend |
| voxtral-tts.c | Voxtral Realtime 4B speech-to-text in pure C |
| vibevoice.cpp | Native port of Microsoft VibeVoice for TTS (voice cloning) and long-form ASR with speaker diarization |
| rf-detr.cpp | Native RF-DETR object detection and instance segmentation |
| locate-anything.cpp | Open-vocabulary object detection and visual grounding (LocateAnything-3B) |
| depth-anything.cpp | Depth Anything 3 monocular metric depth + camera pose estimation |
| face-detect.cpp | Face detection, recognition, demographics and anti-spoofing (SCRFD/ArcFace, YuNet/SFace), replacing the Python insightface backend |
| free-splatter.cpp | Pose-free 3D reconstruction (FreeSplatter): turns a handful of plain photos into 3D Gaussians, no camera poses or GPU required |
| privacy-filter.cpp | Standalone GGML PII/NER token-classification engine powering LocalAI's PII redaction tier |
| LocalVQE | Joint acoustic echo cancellation, noise suppression, and dereverberation |
| local-store | Local-first vector database for embeddings (shipped in-tree) |
We also maintain apex-quant, a per-tensor, per-layer quantization recipe for Mixture-of-Experts models that exploits their structural sparsity to produce GGUFs matching or beating Q8_0 quality - and they run out of the box on stock llama.cpp.
Resources
- Documentation
- LLM fine-tuning guide
- Build from source
- Kubernetes installation
- Integrations & community projects
- Installation video walkthrough
- Media & blog posts
- Examples — including the realtime voice assistant demo (Go client for the Realtime API with tool calling)
Team
LocalAI is maintained by a small team of humans, together with the wider community of contributors.
- Ettore Di Giacinto — original author and project lead
- Richard Palethorpe — maintainer
A huge thank you to everyone who contributes code, reviews PRs, files issues, and helps users in Discord — LocalAI is a community-driven project and wouldn't exist without you. See the full contributors list.
Citation
If you utilize this repository, data in a downstream project, please consider citing it with:
@misc{localai,
author = {Ettore Di Giacinto},
title = {LocalAI: The free, Open source OpenAI alternative},
year = {2023},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/go-skynet/LocalAI}},
Sponsors
Do you find LocalAI useful?
Support the project by becoming a backer or sponsor. Your logo will show up here with a link to your website.
A huge thank you to our generous sponsors who support this project covering CI expenses, and our Sponsor list:
Past sponsors
Individual sponsors
A special thanks to individual sponsors, a full list is on GitHub and buymeacoffee. Special shout out to drikster80 for being generous. Thank you everyone!
Star history
License
LocalAI is a community-driven project created by Ettore Di Giacinto and maintained by the LocalAI team.
MIT - Author Ettore Di Giacinto mudler@localai.io
Acknowledgements
LocalAI couldn't have been built without the help of great software already available from the community. Thank you!
- llama.cpp
- https://github.com/tatsu-lab/stanford_alpaca
- https://github.com/cornelk/llama-go for the initial ideas
- https://github.com/antimatter15/alpaca.cpp
- https://github.com/EdVince/Stable-Diffusion-NCNN
- https://github.com/ggerganov/whisper.cpp
- https://github.com/rhasspy/piper
- exo for the MLX distributed auto-parallel sharding implementation
Contributors
This is a community project, a special thanks to our contributors!