The speech-to-speech path labels its transcription WAV with the session
rate for client commits, so resampling first handed it 16 kHz audio under
a 24 kHz header. Resample only in the chained pipeline, and say so in the
docs.
Assisted-by: Claude:claude-sonnet-5-5 go-test
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
(cherry picked from commit 82cc411af54cf884e4a44a3e96292a323ba5611c)
The committed-utterance WAV is written with a fixed 16 kHz header. The VAD
commit path stores audio that is already resampled to 16 kHz (vadCommit),
but the client-commit path (input_audio_buffer.commit) passed the raw
buffer at the session input rate. For a 24 kHz OpenAI Realtime session the
backend decoded 24 kHz audio as 16 kHz (1.5x speed/pitch): garbage or empty
transcripts, then input_not_committed on the follow-up response.
commitUtterance now resamples from session.InputSampleRate to
localSampleRate, so both paths hand 16 kHz audio to the 16 kHz header.
This replaces the earlier approach of this PR (header at the session rate),
which mislabelled the already-resampled VAD audio (review feedback).
Tests go through the real paths with a 24 kHz session: commitUtterance and
a server_vad commit via vadTick; both must produce a 16 kHz WAV of the
resampled length. The client-commit spec fails without the fix.
Fixes#12747
Signed-off-by: nexxtmobile.de <git@nexxtmobile.de>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
(cherry picked from commit 05ba891da6)
(cherry picked from commit 56e2b6d0cc5cbf1944ef50735f7b499d998eb25c)
Clients written for llama-server, vLLM or SGLang switch template variables
per request with a top-level chat_template_kwargs object, most often
{"enable_thinking": false}. OpenAIRequest had no such field, so it was
dropped: with a Qwen3-style template thinking stayed on, the reply landed
in message.reasoning and content came back empty.
Fold the field's scalar values into the request metadata, so it takes the
path metadata already takes (gRPCPredictOpts, ResolveChatTemplateKwargs).
Booleans become "true"/"false", strings pass through, numbers become their
decimal string; nested values are skipped. An explicit metadata key wins.
Assisted-by: Claude:claude-opus-5-5
Signed-off-by: Cristi An <c4xp@msn.com>
(cherry picked from commit 2cd7425e87)
(cherry picked from commit 5e86ad6bd5286cc8db398d278adaa32e562223ca)
qwen3's chat template wraps <think> in {% if enable_thinking %}, but the
prefill detectors scan the raw template without evaluating the condition.
With streaming and no tools, the synthetic prefill made the Go-side extractor
classify the whole answer as reasoning whenever the client sent
metadata.enable_thinking=false.
Skip prefill detection when the request disables thinking. Real
<think>...</think> blocks that a model emits anyway are still extracted from
the content itself.
Fixes#12939
Assisted-by: ZCode:deepseek-flash
Signed-off-by: xianshujun <xianshupang@gmail.com>
(cherry picked from commit cee7ed4b32)
(cherry picked from commit 29618866d07a05624934bf9f595c95a68c938461)
Signed-off-by: Dedy F. Setyawan <dedyfajars@gmail.com>
(cherry picked from commit 2413f19b7d)
Conflict resolved against current master: keep the strings master already
has for chat.json and add the translations this change supplies for keys
master lacks. Keys added to master after the PR was opened are still
untranslated and fall back to English.
Assisted-by: Claude:claude-sonnet-5-5
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
(cherry picked from commit 8d4799d91b33b2af204e7cc5386d8b984d09f04b)
A model can cite a plausible deep link it never opened, written from
memory. The client cannot tell it from a link a search tool returned,
because server-side MCP tool results never reach it.
evidence_check.urls compares every http(s) link of the final answer,
bare or inside a Markdown link, with the URLs of the request (system
prompt, system, user and tool messages) and of the request's successful
MCP tool results. Links are compared without scheme, fragment, "www."
and trailing slash; a link counts as retrieved when it equals an evidence
URL or is a shorter form of one at a path or query boundary. Unbacked
links (at most five) are listed in an appended note; urls_note replaces
the default note, with {urls} as the placeholder for the list.
Assistant messages are not evidence, so an invented link is marked again
when the client sends it back on the next turn.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Stefan Walcz <stefan.walcz@walcz.de>
(cherry picked from commit 523e8cd3c4)
ChatEndpoint gained a compressor parameter after the spec was written, so
the openai test package no longer compiled.
Assisted-by: Claude:claude-sonnet-5-5 go golangci-lint
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
The comment in sourceURLs named ExecuteMCPToolCall, which the tool
consumer matrix test (core/services/addons/tools) reads as a marker of a
file that discovers or executes tools. evidence.go only reads tool
output, so reword the comment instead of classifying the file.
Docs-Not-Needed: comment-only change, no behaviour change
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Stefan Walcz <stefan.walcz@walcz.de>
(cherry picked from commit 1043d3cbc4)
Keep balanced parentheses in source URLs and decode MCP text arrays before
extracting links. This preserves Wikipedia paths and escaped query
separators instead of appending truncated citations.
Assisted-by: Codex:gpt-6
Signed-off-by: Stefan Walcz <stefan.walcz@walcz.de>
(cherry picked from commit 9b81cb20ab)
A system prompt that tells a model to look things up first does not stop
it from answering from its own knowledge, for example when the question
comes with a picture.
profile.require_tool makes the first model call of a chat request that
offers the profile's server-side MCP tools use tool_choice "required".
Every follow-up call of the server-side tool loop is made with the
client's tool_choice again: a "required" that stayed in place would make
the model call a tool after every tool result and never answer until the
iteration limit.
With a tokenizer template the backend enforces "required" with a grammar
built from the chat template's tool-call format. Where LocalAI builds the
tool-call grammar itself, the first call's grammar leaves out the
no-action function instead.
The first call is not forced when no server-side tool was offered, when
the client set tool_choice or function_call itself, or when the request
does not end with a user message, so a request returning a client tool's
result cannot bounce tool calls between client and server.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Stefan Walcz <stefan.walcz@walcz.de>
(cherry picked from commit 2801938aac)
A research profile asks the model in its system prompt to cite its
sources, but a model can answer from a search result without naming a
single link, and the client never sees the server-side tool results that
held them.
evidence_check.sources appends one line with the distinct http(s) URLs
returned by the request's successful MCP tool calls (at most five, in
order of appearance) when the final answer names no URL itself. An answer
that already links something is left unchanged, and nothing is added when
no tool result holds a URL, so a link is never made up. The label is
configurable through evidence_check.sources_label (default "Sources") for
profiles that answer in another language.
It runs in the existing evidence gate, on the same streaming and
non-streaming paths as the number note.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Stefan Walcz <stefan.walcz@walcz.de>
(cherry picked from commit 4ee6ff1395)
Four recent changes added English strings without the other locales, so
the locale completeness gate failed in the UI unit tests. Add the keys
to the eight locales. The baseline stays as it is.
Assisted-by: Claude:claude-sonnet-5-5 node
A trigger run or an agent job has nobody to approve a step, and so far
the approval mode auto_readonly and the act.* grants were only recorded.
The prompt was the only thing that kept such a run from acting.
Every tool call of an unattended run now goes through one check at the
call. A tool that only reads runs. Any other tool needs the grant of its
class, and the check fails closed:
- act.files for write_file, edit_file, move_file, delete_file and the
built-in bash;
- act.addons, or act.addons:<id>, for add-on tools whose declared kind is
not read, plus act.spend when the add-on gates the tool with a
permission;
- act.spend for MCP tools.
A trigger holds its new acts list within the grants of its agent. A
trigger that declared read only runs, or that a binding in auto_readonly
mode made without asking, is refused every non-read tool whatever it
holds. A binding that cannot be read counts as auto_readonly. A job
holds the act grants of its agent. Sub-agents follow their parent, and a
saved agent cannot be started from an unattended run.
A refused call answers with a plain sentence the model can pass on, not
an error, and writes one audit log line. Triggers set acts through the
REST API; a model never can. A trigger whose plan names a file it writes
keeps act.files.
Assisted-by: Claude:claude-sonnet-5-5 go ginkgo swag
The Tools list gains a group, "What the agent may change", with one
row per builtin grant: its title and a plain-words description, and a
mark on the sensitive ones. The agent form switches them through
useAgentTools; a chat of a saved agent lists the grants the agent
holds, read only. Both draw the group with the same useToolGroups and
SwitchPanel code.
The group has no group setter, so Select all and Disable all, and the
"N of M on" count, leave it alone: a bulk click must not grant a
right.
Assisted-by: Claude:claude-sonnet-5-5 node
The Skills page asks for the skills of the saved agents with
agents=true and shows them with the agent's name. A skill of the
agent layer says who proposed it, can be removed (to the Trash) and
is not offered to other agents. The list of a plain chat or of the
agent form never carries them.
Assisted-by: Claude:claude-sonnet-5-5 go node
A saved agent keeps its own skills in LocalAI/agents/<agent>/skills.
Only a chat that runs that agent sees them. They sit between the folder
and personal layers, so an agent's version of a skill wins for that
agent, and they count as granted without being listed in the agent's
skills.
An agent writes one by a skill change proposal, as before; the new
layer only reads them. A proposal for an agent skill cannot take the
name of a core skill or of a document slot (read from slots.json).
When writing-skills loads in a chat of a saved agent it starts with
"Do not write files. Call propose_change with kind skill.", and the
skills.write grant makes it visible to the agent.
The Skills list gains the agent layer, with the agent's ID and name,
and a skill can be read or removed (to the Trash) with the agent
query parameter.
Assisted-by: Claude:claude-sonnet-5-5 go swag
The draft tools and add-on lists, and the tools list of a stored chat,
read the execution status without an execution scope. The host logged
an error and reported the status as not selected. They now run in a
scope; a draft has no conversation, so its status says "draft".
In the Home composer, switching a built-in tool group on or off reset
the commands state to the loading value and never read it again. The
composer then showed the unknown warning for good. The Hide button also
did nothing for that state, because it needed a binding identity that
the loading value does not have. A draft now keeps the state the list
gave it, and Hide works for any notice that is shown.
Assisted-by: Claude:claude-sonnet-5-5 go ginkgo node
The notice came from navigation state, which survives a refresh, so it
came back after add-ons were on again. A frontend also reports add-ons
as off when the owner does not answer for a moment, which looked the
same as the instance switch being off.
Home now reads /api/features on mount, drops the notice and clears the
history state when add-ons are on, and clears the state on dismiss.
/api/features gains addons_enabled (the instance switch alone), so the
notice can say "not answering right now" with a retry button when the
switch is on and the owner is silent.
Assisted-by: Claude:claude-sonnet-5-5 eslint node go-test
A binding needs a kind now, so the route specs name the platform, and
the spec for an offline binding checks that nothing is kept while no
add-on serves the kind.
Assisted-by: Claude:claude-sonnet-5-5 go test
The add-on guide names one manifest api, one way to move a file and no
migration path, and drops the section on arguments of older tool
versions. The Swagger files are generated again, which also brings in
fields that the last generation missed, and the test of the apps view
now emits every field it documents. The inventory of the add-on suites
lists the specs by their new names.
Assisted-by: Claude:claude-sonnet-5-5 go swag
A connector add-on no longer carries file bytes in the connector
envelope. The ops media and send_file need a transfer handle, and the
gateway refuses a call to them without one, or for a kind that declares
no files.
- The manifest field files.transfer, the inline file of a request and
of an answer (ConnectorFile) and the 1 MiB cap
(MaxConnectorFileBytes) are removed. The 2 MiB limit of a connector
call between hosts drops to 1 MiB, because it holds text and no file.
- A kind that declares no files moves no file: the host answers a sender
who attaches one with a plain sentence and does not offer attach_file.
A direction left out of a declared files block still gets the host
ceiling.
- The inline layer is gone from the file limit, from the API
(inline_bytes) and from the Files section of the connector page.
- The Telegram and Slack add-ons keep only the streaming paths. The
tests moved to the handle path, and the studio test rig gives every
connector add-on a table of handles.
Assisted-by: Claude:claude-sonnet-5-5 go test node
No release carried add-ons, so nothing reads older add-on data or
manifests. Delete the code that did.
- Connectors: the kind and the add-on of the Telegram connectors that
core once carried are gone (LegacyAddon, KindTelegram, the stored
binding normalization, the platform titles). A binding needs a kind,
and its add-on is the one that serves the kind. The conversion of the
old agent pool lists every connector as dropped instead of making a
Telegram binding, and its notes say so.
- The library index reads only the id field of an entry, not addonId.
- A host control that has a capability block (the network policy) is
attached by that block alone, not by a tool that replaces bash.
- The Office create tools no longer answer for the argument names of an
earlier schema, and the identity file note about old addresses is
gone.
- The known debt of the provider-name guard drops the entries of the
conversion and of the resource listing shim, and its ceiling goes
from 13 to 5.
Assisted-by: Claude:claude-sonnet-5-5 go test
No release ever shipped add-ons, so there is no installed base on api 1.
The manifest must now declare api {"min": 2, "max": 2}; a manifest with
api 1, no api or a wider range is refused with a message that names the
range.
Every field that was gated on api 2 is now part of the contract, so the
Min*API constants and their gates are gone, and so are HostAPIMin and
the apiMin host field. Three older-manifest shims go with them: the
resource listing the host implied for one add-on, the legacyArgs hint
for arguments an older version of a tool took, and the tests that
checked the gates. All shipped manifests and the test fixtures declare
api 2.
Assisted-by: Claude:claude-sonnet-5-5 go test
For a kind that declares files.transfer, the connector manager gives the
add-on a handle and not the bytes, in both directions.
Inbound, the add-on streams the file to the host. The chat upload then
reads it from the blob store, so a 6 MB PDF reaches the chat. Outbound,
the manager checks the access of the chat to the file, gives the
process that runs the add-on only the hash and the size of the content,
and the add-on pulls the chunks. The file limit of such a kind has no
inline cap: it is the smallest of the host ceiling, the ability the kind
declares and the limit the owner set for the bot, and the sentence names
the number and the layer. A kind without transfer keeps the 1 MiB inline
cap, and its sentence says so.
Checks: the files of one message may add up to three times the limit;
the media type that the add-on states must be among the types the kind
declares; an outbound file must have a type that the kind declares and
must not start with an HTML page. Every transfer writes one audit line
with direction, binding, size, result, layer and time, and no name or
content. The binding page no longer says "larger files need streaming"
for a kind that streams.
The specs run the real Telegram and Slack add-ons, the real gateway and
the real file operations over scripted platform APIs, with a 6 MB file in
each direction.
Not in this change: a daily byte counter per binding, and a virus scan
(the table has a hook that accepts everything).
Assisted-by: Claude:claude-sonnet-5-5 go test
Echo's RealIP trusted X-Forwarded-For and X-Real-Ip from any peer
when no IPExtractor is set. That fed the login and registration rate
limits and the API trace client address, so a client could dodge a
limit by rotating a forged header.
Add pkg/clientip, one resolver for the whole server. With
--trusted-proxies / LOCALAI_TRUSTED_PROXIES (IPs and CIDRs, empty by
default), the forwarded chain is read only when the TCP peer is
listed. The client is the right-most X-Forwarded-For entry that is
not a trusted proxy; a malformed entry stops the walk. With the flag
unset every request resolves to its TCP peer.
The resolver is set as Echo's IPExtractor and used by the access log
(--log-client), so all consumers agree.
Assisted-by: Claude:claude-sonnet-5-5 go gofmt
One manifest-driven guard replaces the Go-only provider guard. It reads
addons/*/manifest.json and fails when the Go strings of core and pkg,
the React UI source, the locale files or the core docs hold the id of an
add-on with a role (as a whole string or a tool-name prefix) or one of
its declared titles as a whole word.
Third-party mentions (OAuth login, release sources, token formats) and
ordinary words sit in a commented allow-list. The special cases that
remain (propose_change memory kind, the Account memory tab, the Telegram
conversion, the Tables listing shim) sit in knownDebt, each with the item
that removes it, and a ceiling that only goes down.
Reword the sandbox sentences in the container docs and use neutral
sample names in the UI catalog page.
Assisted-by: Claude:claude-sonnet-5-5 go test
Bindings saved before permissions were enforced have no capabilities, so
they deny files from senders, triggers, memory writes and talk-back, and
the owner could only fix that through the REST API.
The shared connector page now has a section for it, so every connector
add-on gets it: a switch and a description for each capability with the
sensitive ones marked, the role matrix (never a sensitive capability, never
above the ceiling), pairing with a code shown once, the approved channels
for talk-back, and a read-only view of the access check with the latest
decisions. A binding with no capability shows a notice that says why the
features are off, with an "Allow the usual" button that turns on the
non-sensitive everyday set.
The kinds endpoint returns the closed list of capabilities with a title,
a description and the sensitive and usual flags. The field is additive.
Assisted-by: Claude:claude-sonnet-5-5 go node playwright swag
The provider guard flags any core string that contains the name of a
shipped add-on that provides a generic role. The web add-on has the id
and title "Web", so the built-in tools group titled "Web" in the tools
picker (and the schema example for it) failed the guard on master.
Rename the group title to "Look things up online", which matches its
"cannot look anything up online" off-state text. Update the schema
example and the Playwright fixtures and specs that read the title.
Assisted-by: Claude:claude-sonnet-5-5 go test eslint
A connector binding gets an approval setting: ask, auto or auto_readonly,
with four limits for the two automatic modes (live triggers, minutes
between runs of a recurring trigger, runs a day, lifetime in days). The
setting is generic, so every connector add-on can show it, and the
shared connector page of the add-ons now draws it. A binding that sets
nothing asks, as before.
This change stores, validates and exposes the setting. The next change
acts on it.
Assisted-by: Claude:claude-sonnet-5-5 go swag node
A saved agent now has a Builtins list: the capabilities its owner gave
it (memory.write, profile.write, skills.write, pages.write,
dashboard.write, triggers.propose, triggers.run, talkback.channels,
talkback.new_channel, files.upload, act.files, act.addons, act.send,
act.spend and the opt-in group saved_agents). The list is empty by
default, and nothing turns a grant on for a template, an import or a
write that omits the field.
A chat of the agent holds the grants the user may use at the turn
(agent scope). A tool exists when the grant is held, the chat did not
switch its group off and the run can hold rights: a sub-agent or a run
that started by itself never does. propose_change and revert_change
now exist only for a chat that holds the grant of at least one kind,
and the schema lists only those kinds.
propose_change gains the kinds skill, page and dashboard. A skill is a
SKILL.md for the personal or the agent layer, and a skill with
locked: true is refused, also as a shadow of a locked personal skill.
A page is a set of files checked against the Pages rules and written
as one change set. The dashboard is a home.json checked against its
schema. All three are re-checked when the owner confirms or edits.
GET /api/chats/agents/builtin-grants lists each grant with its title,
description and sensitive flag. The agent record, the write body and
the export file carry builtins; an old file imports as before and an
unknown name is dropped in an import and refused in a write.
Assisted-by: Claude:claude-sonnet-5-5 go gofmt swag
A connection now has a write mode, chosen on the add-on page: off (the
default), propose (each change waits for the person to confirm) or auto
(a chat of the person makes it at once). The host enforces it for every
write tool of an add-on that holds a credential, before the add-on is
called, and holds the call to the repository list of the connection.
Host:
- Binding.WriteMode, stored in both stores and in the connector API
(write_mode). Absent means off, so existing bindings stay read only.
- A credentialed add-on may declare write tools. They name the resource
with the credential slot argument, never resourceId. A tool can set
approval "always", a ceiling that keeps a proposal even in auto.
- The proposal broker fetches the credential for the preview and again
for the apply, so it is never stored. A mode lowered between the
proposal and the confirmation refuses the apply. A proposal that
awaits a person cannot be confirmed by a chat turn.
- New capability name tools.write for the roles a connector kind
declares.
Add-on, tools (kind write): comment-issue, comment-pr, create-issue,
edit-issue (title, close or reopen), add-labels, review-comment-pr.
They preview plain facts, refuse a stale target, cap and screen the
text for secrets, and keep a receipt journal plus a hidden marker in the
text, so a change whose answer was lost is not sent twice.
Assisted-by: Claude:claude-sonnet-5-5 go gofmt ginkgo node
A connector chat whose sender holds talkback.channels gets the tool
send_message. It writes in the conversation, or as a new top-level
message in a channel that the owner listed on the binding (id, label,
approved; at most 10). The tool names no platform.
An unlisted channel is refused and the error lists the ones that are.
A listed channel that the owner did not approve needs the sensitive
talkback.new_channel for each message, so only a paired sender can use
it. The agent cannot add or approve a channel, so text it read in a page
or a file cannot widen where it writes.
Messages go through the outbox (rate limit and queue), with the size cap
and the mention breaking. Each attempt adds an audit row with channel,
kind, capability, length and a hash, never the text.
A trigger can report to an approved channel with delivery target
channel. The binding, the capability and the listing are checked again
when the run ends; any failure sends the notice to the Inbox.
LOCALAI_DISABLE_CONNECTOR_TALKBACK turns both off.
Assisted-by: Claude:claude-sonnet-5-5 go swag
ASR backends like Nemotron-3.5 append end-of-utterance language tags to
transcripts ("<de-DE>", "<en-US>", ...). The committed transcript feeds
the LLM conversation (the user item Transcript field), but the per-utterance
tag is unreliable for short/ambiguous utterances (a German "Hallo." tagged
<en-US>) and the LLM treats it as a language signal, switching the reply
language against the session instructions.
commitUtteranceWithTranscript now strips the EOU tags from the transcript
that goes to appendUserItem/generateResponse (the LLM context). The emitted
input_audio_transcription event keeps the raw transcript (debugging /
client display).
Regression tests: stripEOULanguageTags unit specs + a client-commit spec
asserting the emitted event keeps the raw tag while the LLM user message is
tag-free.
Signed-off-by: nexxtmobile.de <git@nexxtmobile.de>
(cherry picked from commit dff488044c)
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Anthropic sends tool_choice as an object: {"type": "auto" | "any" |
"none"} or {"type": "tool", "name": ...}. The endpoint only read the
"tool" object, so {"type": "none"} still offered the tools and
{"type": "any"} did not require a call. The object was also forwarded
to the backend unchanged, and the llama.cpp backend turns any object
tool_choice into "required", so even {"type": "auto"} forced a tool
call there. Read the object form, and forward the OpenAI equivalent.
Assisted-by: Claude:claude-opus-5-5
Signed-off-by: jaeminbird <jaeminbird@gmail.com>
(cherry picked from commit 5bd07865a3)
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
/v1/messages always answered stop_reason "end_turn" for text replies,
even when generation stopped at max_tokens. Clients that continue a
truncated turn (the Anthropic SDKs, Claude Code) treated the cut-off
reply as complete. Use the same check the OpenAI endpoint uses for
finish_reason "length", in both the streaming and non-streaming paths.
Assisted-by: Claude:claude-opus-5-5
Signed-off-by: jaeminbird <jaeminbird@gmail.com>
(cherry picked from commit d905ab5fdc)
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
The finetune.start and quantize.start operations defaulted to the
backend names "trl" and "llama-cpp-quantization", and the backend
lister had one method per engine family. Both were special cases that
a new engine could not use.
The lister now answers for a gallery tag. A start call that names no
backend gets the first installed backend of the tag in gallery order,
and a named backend must carry the tag. With no engine installed the
call fails with a 400 that says so. The shipped gallery order keeps
trl and llama-cpp-quantization as the defaults, and a test pins it.
The change is additive for add-ons: no manifest field or operation
changes.
Assisted-by: Claude:claude-sonnet-5-5 go gofmt
The GitHub add-on made "GitHub" a provider name, so the test flagged three
strings that name the service itself: the OAuth sign-in button, the footer
link to the project repository and the "GitHub releases" package source in
the egress registries hint. None of them names the add-on, so add narrow,
reasoned allow-list entries for exactly those keys.
Assisted-by: Claude:claude-sonnet-5-5 node npm
The previous commit was made on a stale base and so reverted work that
landed on master before it. This commit puts every file back to the
state of its parent's parent. The permission enforcement of the
previous commit stays.
Assisted-by: Claude:claude-sonnet-5-5 go
Operators could not tell which client sends requests that are refused
with 401. Add --log-client (LOCALAI_LOG_CLIENT, off by default). When
set, the HTTP request log line also carries duration, remote, user_agent
and, when present, request_id.
The address is the TCP peer. It is never read from X-Forwarded-For,
X-Real-Ip or Forwarded, because LocalAI has no trusted-proxy setting and
any client can forge those headers. Behind a reverse proxy the value is
the proxy address. Authorization headers, cookies, query strings and
bodies are never logged. User agent and request id are cut to a fixed
length and have control characters replaced.
Assisted-by: Claude:claude-sonnet-5-5 go gofmt
The capabilities, role grants and pairing that a connector binding
stores now decide what a sender can make the agent do.
Authorize is one pure function: the capability must be in the owner's
ceiling for the binding and the sender must be on the allowlist. A
sensitive capability then needs a paired sender; roles never count for
it. Any other capability needs a pairing or a role that the owner
granted and that the kind lists. The default is no.
Each message gets a grant for its sender. A message that joins a
running turn narrows it to what both senders may do. The tool builder
offers a gated tool only for a held capability, and the tool checks
again when the model calls it, so a call that gets through is refused
with the reason. Today this covers propose_change (memory and profile
kinds), revert_change, the trigger tools, and files sent to the bot. A
chat worker gets the grant with the start request.
The capability gate that the trigger features ask is now answered by
Authorize, and the application sets it. Other features can add a
global switch with AddCapabilitySwitch.
Every use records an allow or deny row (time, binding, sender IDs,
roles, capability, result; no text or codes). GET
/api/chats/connectors/{id}/access returns the effective permissions
per kind of sender and the latest rows, for an add-on settings page.
A binding saved before this change has no capabilities, so its senders
lose these features until the owner enables them.
Assisted-by: Claude:claude-sonnet-5-5 go swag
The agent form offered a different set of choices than the chat: grants
as checkboxes for add-ons, connectors and skills, a free-text list of MCP
servers, and nothing for the built-in tool groups, so a person could not
tell which tools an agent can use.
The form now draws the lists of the Tools menu with the same components
(SwitchPanel, SkillsPanel) and the same rows (useToolGroups), fed by
useAgentTools: the draft lists of a new chat on the agent's model, with
the agent's grants and switches applied. A switch changes the form, and
the agent is saved with it:
- built-in groups: builtins_off and builtins_on (an opt-in group such
as saved agents starts on only when listed);
- add-ons, connectors, skills: the existing grant lists, so what is
switched on is what is granted. A granted entry this account cannot
see stays in the list, marked, so it can be removed.
The network choice and the folder and mode choice stay where they were.
The MCP servers of the model stay a free-text list in their own fold,
since a model that is not chosen yet has none to list.
Assisted-by: Claude:claude-sonnet-5-5 eslint playwright
The Tools menu built its rows inside the composer, drew only some of
them with a description, and showed a group it had no words for ("changes",
the memory and profile edits) by its raw ID.
The rows now come from useToolGroups, and every row has the same three
parts: a title, a description of what the model can do with the switch on
(and, once it is off, what it loses), and a state line (who serves the
tools, whether commands are isolated, why a switch cannot be used). The
words come from the server, which declares a title and a description for
each group and each add-on; a translation by group ID wins when the
locale has one, so a group added later needs no change here. An add-on
row shows the title its manifest declares instead of its ID.
SwitchMenu is split into the chip and SwitchPanel, and SkillsMenu into
the chip and SkillsPanel, so another page can draw the same lists inline.
New strings are in every locale.
Assisted-by: Claude:claude-sonnet-5-5 eslint playwright
The Tools menu of a chat listed a ninth built-in group, "changes", by its
raw ID with no text, and only some rows said what a switch did. The
registry now carries a title, a description of what the model can do when
the group is on, and what happens when it is off, for every group. Groups
that nobody described fall back to the label and description of their
first tool, so a new tool group needs no change in the interface.
An add-on row can show the title and description its manifest declares:
the broker may implement the optional AddonLister, and the lists carry
both fields next to the ID.
A saved agent gets builtins_off and builtins_on. builtins_off is a
ceiling: a chat of the agent shows the group off with a reason and cannot
switch it on, and a sub-agent follows it. builtins_on lists the opt-in
groups the agent's chats start with. Both are additive columns, a write
that does not name them keeps the stored ones, and groups the server does
not offer are dropped so an imported file still loads. An agent with
switches exports as version 3, so an older server refuses the file
instead of giving the agent every tool; a single-folder agent without
switches still exports as version 2.
Assisted-by: Claude:claude-sonnet-5-5 go gofmt swag
Add a manifest credentials declaration, host-enforced per-binding
resource lists, Request.Credential and ResourceKey set by the host, and
a typed ErrRateLimited that reaches the model as rate_limited.
Assisted-by: Claude:claude-sonnet-5-5 [go]
The Files section of the agent editor is now a list of rows. Each row
has a folder picker, a read or write switch and a remove button. An
Add folder button stops at eight. A folder inside another one shows the
mode of the outer folder, because that is the mode it works at. The
agent chips, the Tools menu line and the Files page marks read the whole
list. The strings are in all nine locales.
Assisted-by: Claude:claude-sonnet-5-5 eslint playwright
POST and PUT on /api/chats/agents take folders, a list of path and mode
rows, up to eight. The old folder field still works and is read as a
list with one entry. Responses carry folders, the effective mode of each
row and a notice with the count of shared folders a chat does not mount.
A folder inside another folder of the same agent takes the mode of the
outer one, because the file store gives a node the highest mode among
the roots above it and cannot hold a read only folder inside a write
folder. The editor data shows that mode.
An agent with one folder still exports as version 2. An agent with
several exports as version 3 with folders. A server that reads version 2
refuses version 3. Import keeps a folder that is not found as a row
that is not chosen.
Assisted-by: Claude:claude-sonnet-5-5 go gofmt
A chat that does not exist yet has no execution status to read, so its
isolation is unknown by construction. Three places treated that as a
failure:
- Switching an add-on in a draft reset the commands state and never
refilled it, so the "isolation is unknown" notice stayed. The draft
now keeps the state its list gave it.
- The skills list of a draft hid the document skills of the selected
commands provider, because they need isolated commands. A draft list
now offers them; the turn of the real chat still gates them on the
status it reads. The draft add-on switch also reads the skills again.
- The shell row of the Tools menu printed the generic unknown text for
the draft state. It now says nothing about isolation there.
In a real chat, switching an add-on or a tool group now reads the
chat's network view again, so the network controls appear once the
provider serves commands.
Assisted-by: Claude:claude-sonnet-5-5 go gofmt playwright
A connector chat can now propose a trigger, the person answers it in the
chat, and the run reports to the chat.
The trigger tools propose_trigger and list_triggers are offered in a
connector chat when a capability gate allows triggers.propose for the
binding. The gate is a small interface on the service with a deny-all
default, so nothing changes until the application sets one. Change and
delete stay in Studio.
A proposal from a chat stores the binding, chat and thread, a four
character code and a 24 hour expiry. The host writes the proposal and the
code in that chat. "confirm CODE" or "cancel CODE", as a whole message in
the same chat and thread, is handled before any model turn and needs
proposals.confirm. A binding may make 3 proposals an hour. Proposals that
lapse are marked expired, and the proposal card shows that state.
A trigger carries a delivery: the Inbox (default) or the chat of the
proposal, with the modes always, on_change and on_problem. The host writes
the last message of the run, cut and with broadcast mentions broken; a
message of exactly [SILENT] writes nothing; three failures in a row raise
one alert per 6 hours. A failed delivery falls back to the Inbox. Host
messages of a binding go through an outbox of 20 a minute that queues the
rest. While triggers.run is off, runs of such a trigger are skipped, not
replayed, and the record is unchanged. The audit records each step with a
hash and no text.
Assisted-by: Claude:claude-sonnet-5-5 go test ginkgo eslint
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Ettore Di Giacinto