llama-swap: baixar versions anteriors

This release includes security dependency updates and requested quality of life changes from users:
- the UI no longer expects to be run under a root path making it easier to serve llama-swap behind another proxy
- the max request size is configurable now. Users should be free to OOM as much as they want. llama-swap ain't your mom!
- allow manual override of values in /v1/audio/voices. This is a non-standard endpoint but makes working with upstream audio engines a bit nicer
Details below:
- PR #1224 cmd/kubeswap: attach DRA ResourceClaims and mount NFS volumes: add repeatable --resource-claim and NFS --volume support so Kubernetes backends can share a DRA GPU claim and mount models without a static PV and PVC for each export (#1222) by @gabrielcosi
- PR #1225 Add voices override for speech models: let local models configure capabilities.voices so /v1/audio/voices returns a reliable list without starting the model or querying the upstream server (#1207) by @mostlygeek
- PR #1228 ui,server: serve the UI under a reverse proxy base path: use relative asset, API and redirect paths so the UI works behind a prefix-stripping reverse proxy without routing its endpoints at the proxy root (#1227) by @mostlygeek
- PR #1249 internal/server: make the request body size limit configurable: add maxRequestBodyMB for model endpoints and Tailcat, replacing the fixed 250 MB cap while retaining 250 MB as the default (#1245) by @mostlygeek
- PR #1257 AGENTS.md, go.mod: use go 1.27.3, trim AGENT rules: raise the Go version to 1.27.3, update dependencies and simplify contributor instructions by @mostlygeek

A couple of small improvements:
- support serving llama-swap behind a path on a reverse proxy. Helpful if you want to front llama-swap with nginx and serve it under a specific path.
- add support for manually defining voices for text to speech models
- a small kubeswap update
- PR #1228 ui,server: serve the UI under a reverse proxy base path: makes the browser side base-path agnostic (relative asset, redirect, favicon and PWA URLs plus an apiUrl() helper that re-attaches a stripped prefix) so llama-swap works behind a plain prefix-stripping proxy at https://host/subpath/llama-swap/ without routing /ui, /api and /v1 at the proxy root; includes an nginx and Caddy guide (#1227) by @mostlygeek
- PR #1225 Add voices override for speech models: adds a models.capabilities.voices capability listing the voice ids an upstream server supports, because the non-standard GET /v1/audio/voices endpoint returns unreliable responses; falls back to probing that endpoint on a best-effort basis when the list is not set (#1207) by @mostlygeek
- PR #1224 cmd/kubeswap: attach DRA ResourceClaims and mount NFS volumes: adds a repeatable --resource-claim name=claim flag so backends can get GPUs via the NVIDIA DRA driver (kubeswap only references the claim, pods sharing a claim share its device) and an nfs:server/export:path[:ro] kind for --volume so models can be mounted from NFS without a static PV/PVC per export (#1222) by @gabrielcosi

This release includes a new auto:orphans set for swap matrix rules. This new set automatically includes all models that were not explictly added to a set. It is defined as (orphan1 | orphan2 | orphan3 | ...). The orphans set is handy for definitions like: +always & +auto:orphans, where +always is a set of models that are always loaded and should never be swapped out.
The /audiocpp endpoint was added to make working with audio.cpp server eaiser. The audiocpp_auto model is automatically used when the endpoint is requested. audio.cpp's OpenAI compatible endpoints can be accessed directly via /audiocpp/v1/audio/*. Very handy when used with llama-swap's unified docker container which builds audio.cpp with the UI and native model downloader.
Also one tiny commit to improve compatibility with llama-server's decision model functionality.
Thanks to the contributors in this release, especially @knguyen298 for being patient and keeping his PR updated.
- PR #1026 internal/matrix: add synthetic auto:orphans set: adds a +auto:orphans selector that expands to every model not named in another set, so a catch-all set like "+always & +auto:orphans" picks up new models automatically (#1025) by @knguyen298
- PR #1223 Add /audiocpp endpoint for audio.cpp UI passthrough: serves audio.cpp's web UI and model management through a dedicated /audiocpp/ endpoint bound to the reserved audiocpp_auto model, like /comfyui/, so the unified container works without naming a model per request by @mostlygeek
- PR #1220 internal/capcompat: detect decision models from llama-server's listing: fix decision models being reported as out: [text] (and capabilities.out: [decisions] failing validation) because discovery only read /props; output_modalities from /v1/models is now used, and tools are only advertised when the model outputs text (#1219) by @nazeshinjite

I've been busy, busy, busy and have not had as much time for llama-swap as I
would like. This release includes reliability improvements, improved
compatibility and some small quality of life improvements. Thank you for the
amazing contributors for this release! Everyone who files issues or submits a
PR is greatly appreciated.
- PR #1213 Add confirmation UI and progress tracking for WoL wake: wol-proxy gets a -require-confirm flag that shows a "Start server" page instead of waking on every request, plus a retro loading page with a progress bar estimated from the last wake duration by @mostlygeek
- PR #1215 internal/router: strip browser headers before proxying to peers: fix UI chats failing with a 401 against an api.anthropic.com peer, which treated the forwarded browser Origin header as a direct browser call; Origin, Referer, Sec-Fetch-* and Access-Control-Request-* are now removed (#1214) by @mostlygeek
- PR #1212 internal/process: report loading progress from the health check: parse progress/message from not-ready health check JSON bodies, show it on the Models and model detail pages, and restart the healthCheckTimeout countdown when progress changes so long loads don't time out (#1079, #1208) by @mostlygeek
- PR #1196 internal/perf: restart loop to trigger GPU collector when the monitor still running: fix GPU stats stopping for good after nvidia-smi exited; the exit error is now logged and the collector restarts with a 5 to 30 second backoff (#1155) by @hpdkhoa
- PR #1211 internal/process: add a per-model disableKeepAlives option (#1205): fix intermittent 502 proxy error: EOF when a POST reused a pooled connection that llama-server closed after a streamed response; disableKeepAlives: true opens a new connection per request by @johnnieCR
- PR #1199 fix(metrics): show oMLX token rates in Activity: read oMLX's usage.prompt_tokens_per_second and usage.generation_tokens_per_second as a fallback so rates no longer show as unknown by @junmo-kim
- PR #1198 internal/capcompat: add Gufo inference engine prober: add a default capability prober that uses only /v1/models data as the fallback for servers without a custom prober (vLLM, Gufo, SGLang, etc.), replacing the vLLM-specific one by @berney

systemone models are mega hyped right now. This release adds support for
/v1/systemone for that fast thinking goodness(?). Also requests missing a
Content-Type will get one automatically added for the inference endpoints.
- PR #1197 Add /v1/systemone endpoint to POST JSON routes: route llama.cpp's new /v1/systemone decision model endpoint to the model named in the JSON body (#1195) by @mostlygeek
- PR #1194 internal/server: accept JSON bodies with a missing or wrong Content-Type: fix a misleading "no model id could be identified" 404 when curl -d sent JSON as form-urlencoded; a middleware now relabels JSON object bodies as application/json and returns a 400 naming the expected format for unusable bodies (#1192) by @mostlygeek

This release has a bunch of small compatibility improvements requested by the community.
Also the changelog summaries have been changed to only AI for the list below. These notes at
the top will be written by humans (or my dog) for humans. The AI's writing more correct but
fuck it we have too much AI slop in our lives already.
- PR #1189 Increase ComfyUI concurrency limit from 50 to 999: fix ComfyUI stuck on the splash screen via /comfyui because one tab makes more parallel requests than the old minimum of 50 allowed (#1188) by @mostlygeek
- PR #1187 docker/unified: compile sm_70 in the CUDA 12 image for V100: fix ik-llama-server aborting on V100 GPUs, where only sm_61 PTX without a WMMA flash-attention kernel was available (#1185) by @mostlygeek
- PR #1158 Hardware Detection: detect Intel GPUs on Linux via xpu-smi, including dedicated Arc memory, and read the Metal version on macOS across system_profiler key names by @anantshri
- PR #1183 docs/kb: document that tailcat.models accepts selectors: note that aliases, selectors, profile pins and peer model names are accepted, with a selector example by @mostlygeek
- PR #1182 Build audio.cpp with static eSpeak-ng and install phoneme data: link eSpeak-ng statically and ship its phoneme data so Kokoro, SanoTTS and Inflect phonemization works without a writable cache (#1181) by @mostlygeek

This release fixes a crash at startup on macOS 27 with Apple M6 hardware
(#1178). Hardware detection reads the CPU frequency through gopsutil, which
expects a voltage-states5-sram property that is missing on M6. It passed the
empty value straight to CFDataGetLength, which segfaulted before the server
could start listening.
As a temporary fix, gopsutil now points at a fork with the patch from
shirou/gopsutil#2163, which checks for the missing property. This will be
removed once the upstream project picks up the patch. Some Go dependencies
were also updated as required by the newer gopsutil.
- PR #1180 go.mod: use gopsutil PR #2163 to fix darwin/arm64 crash: replace gopsutil with a patched fork that guards the CPU frequency probe against a missing property on Apple M6

Logging is split into three separate streams: proxy, upstream and HTTP access
logs. HTTP request lines used to be mixed in with the proxy log, which made it
hard to search. They now have their own stream and their own tab in the log
viewer. The logToStdout setting accepts a comma separated list of streams, such
as "proxy,http", and the old values still work.
Upgrading: log data has moved off GET /api/events to a new GET /api/events/logs
endpoint. Pick streams with ?stream=proxy&stream=http. Tools that read logs
from /api/events need to switch to the new endpoint.
The UI sidebar can be resized by dragging its edge, and the width is
remembered. The model page has a new header that shows the model name, a
copyable ID, its state and how long it has been ready. The API also returns a
new readySince field for models that are ready.
The Hardware page now recognizes DGX Spark (GB10) systems. Before this, it
could not count sockets on arm64, labeled the unified memory as dedicated
memory it could not detect, showed only the efficiency cores and had no power
figure. It now shows the system model, the hybrid CPU layout, shared memory
and the nominal 140 W SoC power.
The "install as app" prompt now works when apiKeys is set. The browser fetches
the web app manifest without an API key, so it got a 401. The manifest is now
served without auth, like the favicon.
The release process now checks for Claude or Codex before generating the
changelog and uses a newer model for it.
- PR #1177 internal/server: serve web app manifest without requiring auth: let browsers fetch site.webmanifest when apiKeys is set so the install prompt appears
- AGENTS.md: revise contribution guidelines for PRs: agents check pull requests and issues against CONTRIBUTING.md and ask when there is a problem
- PR #1172 Refactor logging to support multiple streams and separate log endpoints: split proxy, upstream and HTTP logs, add /api/events/logs and allow a stream list in logToStdout
- PR #1163 hw: detect DGX Spark (GB10) systems: report the system model, hybrid CPU, sockets, shared memory and nominal power on GB10 boxes by @lemassykoi
- PR #1173 ui: sidebar resize, model uptime tracking, small tweaks: drag to resize the sidebar and show how long each model has been ready
- scripts/add-changelog.sh: use gpt-6-sol: update the model used to generate changelog entries with Codex
- Makefile: detect harness before release: use Claude or Codex to generate the changelog and stop with an error when neither is installed

Tailcat is updated to v0.7.0 and now checks allowed clients in the HTTP
handler. A client that is not allowed gets a 403 instead of a TCP timeout,
which makes peer connections much easier to troubleshoot, and changes to
tailcat.allow now apply on hot reload. A bug that dropped the connection when
a node key was not recognized, which could leave clients hanging, now also
returns a 403. Peer connections recover better when the server disconnects,
and they wait longer for the first response header by default.
Upgrading: an empty tailcat.allow list now denies every client. To keep
allowing anyone to connect, set tailcat.allow to "*".
The output of cmdStop goes to the process log again, the same as cmd. This was
lost in an earlier change (#790). A cmdStop that leaves a background child
holding its output open can no longer hang the stop, and hitting that wait
limit after a successful stop is logged as a warning instead of a failure.
The Models page shows a status icon for each target of the selected profile,
and those models can be loaded and unloaded from the profile pane.
The unified Docker image adds mp3 support and native downloads to audio.cpp.
The image build also stops failing on untagged main-branch commits. Those
builds fell back to looking up the "latest" release through the GitHub API,
which rate limited the shared CI runners. The installers now build the commit
directly and fall back to git tags when the API is not available.
Release notes on GitHub now use the CHANGELOG.md entry for the tag, and
changelog entries are written as plain-language summaries that credit outside
contributors.
- release: improve generated changelog and release notes: generate readable changelog entries, credit outside contributors and use each entry as the GitHub release notes
- PR #1169 internal/tailcat: various reliability and bug fixes: return a 403 for unrecognized node keys, improve peer reconnects and raise the default time to first header
- PR #1170 ui: link profile targets and add load controls: show status and load/unload controls for profile target models by @sousekd
- PR #1165 internal/process: log cmdStop stdout and stderr: send cmdStop output to the process log and stop a lingering child from hanging it by @dividehex
- PR #1166 internal/tailcat: upgrade to tailcat v0.7.0 and use Server.Listen: check tailcat.allow per request, deny by default when empty and add a "*" wildcard
- PR #1157 docker/unified: audio.cpp add download and mp3 support: add mp3 and native download support
- PR #1156 docker/unified: fix vllm-wrapper build on untagged commits: build untagged commits directly and fall back to git tags when the GitHub API is rate limited

Changelog
- f00d375 changelog: v257
- 6f17d9e ui/playground: add a searchable model picker (#1153)
- eeac3e6 internal/capcompat: automatic context and modality discovery
- 769dbac internal/server,internal/config: Add tighter CORS controls
- 29d10df docs,server: improve comfyui endpoint and add a guide (#1145)
- 96e6f94 internal/store: preserve request totals after pruning (#1136)
- 6c36b88 internal/server: fix CORS handler returning too early
- 2edad4a internal/server: support proxy forwarded for headers
- 76ab83b cmd/kubeswap/chart: fix the default demo's loopback bind (#1131)
- 16c026c .github/workflows: fix helm chat publishing (#1129)
- c5753ce cmd/kubeswap: kubernetes backend for llama-swap (wrapper + helm chart) (#1114)

Maintainer note: This is the v256 Big-F-Release. I've been very busy and many small improvements have been squatting in main waiting to be unleashed. A few highlights:
Playground Chat UI improvements
- tightened up the margins and cleaned up the design to be more mobile friendly
- more timings and stats than you can shake a stick at
- improved system prompt and added more hyper param controls (do who uses these anymore?)
Thanks to contributors:
- JoC - Started the model idle-time window after models become ready or restart PR #1095.
- Kris - Added live generation statistics to Chat PR #1099.

Changelog
- 6701d0d changelog: v256
- 2409272 CHANGELOG.md,docs: fix pull request link
- b98a8ff changelog: add scripts and automation to generate changelog entries (#1125)
- 40b36d9 ui: make the playground chat responsive and lighter (#1120)
- 0e1f797 internal/server: reduce zstd pool memory consumption (#1123)
- aecf92e internal/store: split repository interfaces from the sqlite backend (#1122)
- 21bc145 process: start TTL idle window when model becomes ready (#1095)
- 41ec321 internal/server: add global concurrency semaphore (#1110)
- e1526c8 cmd/vllm-wrapper: use default timeouts
- 8fa8589 internal/server: add tabbyapi usage data extraction (#1104)
- 1686326 ui: show chat stats in Help agent
- acb7759 ui: show live generation stats in Chat (#1099)
- 6b18115 ui: keep Work status bar stable while streaming (#1101)

Changelog
- 7761aa1 tailcat: update to v0.6.0 (#1098)

Changelog
- 6c7f958 docker/unified: add cuda13 with multi-platform support for arm64 (#1093)
- 82ccfbf filters: set-if-undefined params via "?" key suffix (#1075)
- 41f1431 ui/tailcat: mask connection token, add shareable peer config (#1091)
- 8365956 ui/help: raise Docs Agent max_tokens from 4096 to 65536 (#1089)

Changelog
- afba24b ui/help: refine Help agent topics and controls (#1088)
- a77f107 Replace config path argument with jq query support (#1087)
- 8a06d56 tailcat: add private server and peer connectivity (#1074)
- 7a14664 docker/unified: make CUDA version and architectures configurable (#1072)

Changelog
- e31a1ad Split Docker build into per-project stages for CI (#1071)
- e2d6f93 ui: improve Help page (#1070)
- 6ffd67f internal/reference: add indexed docs and /api/tools endpoints (#1054)
- 32c3d66 config,server: add startup profile hook (#1053)
- 1fc0490 expose context_window on models endpoint (#1063)
- 51d1969 cmd/wol-proxy: add -api-key option for use in health checks (#1055)
- a6f5ae3 cmd/wol-proxy: add liveness check (#1057)
- 8a32e05 Typo (#1049)
- 4cd0247 config: accept video as an input and output modality (#1048)
- bfbb99c cli: add -validate flag to check config and exit (#1046)
- 9b4b3a0 ui: ignore IME composition Enter in keydown handlers (#1045)

Changelog
- 4ec3175 internal: record client disconnects as 499, not 200 or 502 (#1040)
- 23432e6 internal/server: rate limit sendBuffer full warnings (#1043)
- 0bdb372 internal/server: parse vLLM speculative decoding metrics (#1039)
- cb53616 internal/swaputil: OpenAI-compatible error bodies (#1038)
- 23c28fe docker/unified: build audio.cpp as a deployment build (#1023)
- b40d20e ui: render ANSI colors in logs (#1019)
- 08f1076 ui-svelte: Add markdown activity export (#1020)

Changelog
- 60226b6 docker/unified: add llama-bench, vllm-wrapper, audio.cpp (#1011)
- 44a17d7 README.md: reorder list of features
- 3bccbef Support argv-based vLLM startup in vllm-wrapper (#979)
- be683dc ui-svelte,internal/server: show capability tags on Models page (#1007)
- 2edfa73 api: support /v1/task/run for audio.cpp (#982)
- be50598 AGENTS.md: tweak rules around pull requests
- 3e583ad CONTRIBUTING.md: update rules
- 038ae45 expose meta.n_ctx on models endpoint (#984)
- 6333ee3 router: add /models endpoint (#983)

Changelog
- f94c94a various: improvements to code layout and naming (#1003)
- 249797f internal/server: add ComfyUI compatibility endpoint (#1002)

Changelog
- 00515cd internal/server,shared: preserve percent encoded in upstream (#988)

Changelog
- 40027d6 ui-svelte: security update
- 298848d internal/hw: detect inference host hardware (#978)

Changelog
- 22df230 ui-svelte: show configured selectors and strategies on Models page (#975)
- 8d61908 internal/perf: fix rocm-smi GPU memory utilization (#973)
- 1f3c68e docker: install rocm-smi for vulkan backend (#968)

Changelog
- 30470a4 ui-svelte: show profiles on model page (#966)
- 27782ab AGENTS.md,CONTRIBUTING.md: update contribution guidelines (#964)
- b9c3738 docker: add -watch-config (#963)

Changelog
- 0c42333 internal/matrix: solve matrix expressions symbolically (#960)

Changelog
- 4c50064 internal/config: relax matrix model reference constraints (#957)
- 4bffcfa cmd/vllm-wrapper: add helper for sleep/wake (#941)
- fa607fe internal/config: refactor macro expansion (#955)

Changelog
- 7aa7f52 internal/server,ui-svelte: add peer model namespaces (#950)
- 5e9f887 Fix process start race during concurrent stop operations (#949)
- 12af351 ui-svelte: remove audio file size limit (#948)

Changelog
- 8b61e3d internal/server: implement selectors (#942)
- c6adf57 internal/server,shared: reduce request model body functions to one (#940)
- 3e87c02 internal/server: add support for profiles (#935)
- 378a26e ui-svelte: add support for different v1/audio/voices response formats (#932)
- fc8b29f docker: add FFmpeg support for whisper.cpp (#785)

Changelog
- 6b5320d internal/server: expand inflight request details (#923)
- cd22a8c Add configurable UnloadTimeout variable (#904)
- 61f83ac ui-svelte: code-split routes and trim highlight.js/chart.js bundles (#925)
- 7462973 .coderabbit.yaml: disable annoying unit test creation
- 799eedb internal/config: preserve yaml anchors in capabilities (#918)

Changelog
- dd81801 internal/config: preserve yaml anchors in capabilities

Changelog
- 3c79461 internal/server: improve vllm response metric calculation (#913)
- 8945d2b internal/config: resolve macros in capabilities fields (#907)

Changelog
- b834706 internal/server: add status to v1/models for llama-server (#901)
- 6291bb8 internal/store: persist activity metrics to sqlite (#898)
- 579b1e7 fix starhistory

Changelog
- c59816b internal/server: show inflight activity requests (#895)
- d8ed077 Add llama-tts binary (#894)
- 0d2d851 Makefile: improve build version stamping (#891)

Changelog
- 4a6b8a8 internal/router: reject concurrency excess before streaming (#889)
- 3023ab4 add /props to modelGetRoutes (#886)

Bug and Performance improvements for the UI
- fixed streaming logs blocking llama-server inference. Logs are pushed to connected UI clients via an async channel now so if they are slow it does not affect reading from stdout of llama-server
- replaced the placeholder.txt with a go build tag approach - fix CI
Changelog
- 083e421 internal/logmon: decouple log broadcast from Write (#878)
- a667c57 internal/server: gate UI embed behind embed_ui build tag (#880)

Changelog
- 25b27cb ui-svelte: add rounded borders
- 18cc854 update hero image

This release adds the -config-dir flag! Using a config-dir it’s now possible to split huge config.yaml files into fragments that get automatically merged together.
You can have one model per file and combined with -watch-config models can be enabled or disable by moving them in or out of -config-dir.
The -config and -config-dir flags are compatible and can be used at the same time or exclusively.
Changelog
- 32bc781 internal/config,watcher: add -config-dir (#873)

Changelog
- 316ad63 config,server: add upstream.ignorePaths (#869)
- e37077a feat: hide performance menu item if disabled (#832)
- eff9b60 server: capture failed (non-200) LLM requests (#862)
- 9bcddad internal/server,ui: add new Acitivty page column - Drafted (#859)

Changelog
- a15e479 proxy: meter /upstream requests via metrics middleware (#858)

Changelog
- ed77385 ui: improve manual model load and cancel (#847)

Changelog
- 0cfe5a6 Makefile,internal: fix websocket regression and other small things (#830)
- 44e1501 internal/process,server: fix unload regression (#828)
- 46cea36 proxy: remove legacy code. Thanks champ 🫡 (#822)
- ccfba0d docker: fix arm64 cpu image downloading amd64 llama-swap binary (#819)
- ddfae90 Change cron schedule for container builds

Changelog
- 29d3d9b perf: add macOS GPU monitoring via mactop and ioreg (#816)

Changelog
- 9be9a87 internal/process: improve windows shutdown behaviour (#808)

Changelog
- 4ca9c47 Makefile,internal/server: various release tweaks
- 146a9ea ui-svelte: update build directory (#801)

Changelog
- 02e015f Introduce new routing backend (#790)
- 63bc266 Add new power draw column header for rocm-smi monitoring (#788)

Changelog
- 636b53e Improve rocm-smi performance monitoring (#775)
- 59cd3b6 Added Windows performance monitoring using nvidia-smi (#773)
- 5d1e62d Disable auto review feature in coderabbit config
- dbb869d Increase inactivity thresholds for stale issues
- 26bb17e config.example.yaml: Improve matrix vs groups info

Changelog
- 2982dd3 ui-svelte: update link to performance discussion thread

Changelog
- b2fcc2d ui-svelte: fix cached tokens total counting -1 sentinel (#760)
- 6a9c4ef fix: use --loop instead of -loop for nvidia-smi (driver 540+ compat) (#759)

Changelog
- 0c813e4 ui-svelte: package updates
- fe71e8a proxy,ui-svelte: improve support for v1/messages and v1/responses (#758)

Changelog
- c79114d proxy: fix logger not checking matrix for processes

Changelog
- 430166d proxy: fix zero duration for non streaming responses (#723)
- 5b4beac fix: ?no-history flag and improve /logs monitoring docs (#721)

Changelog
- fd3c28f Refactor Activity Page (#710)
- a846c4f config: remove hard cap on macro length (#718)
- 5bae33a ui-svelte: default theme to user preferred color scheme (#712)
- 8f4ff01 ui-svelte: make it easier to toggle panels in logs view

Changelog
- e8d4384 ui-svelte: support reasoning and reasoning_content (#708)
- ce28485 ui-svelte: add prompt processing histogram (#705)
- 3cd7837 fix: support architecture-specific download URLs in install script (#698)

Changelog
- 0b31cca ui-svelte: fix histogram calculation (#695)
- 5938dbe Push unified docker images on scheduled runs (#694)

Changelog
- 66639e8 proxy: replace fsnotify with stat-poll watcher and add SIGHUP reload (#685)
- 625b296 docker/unified: add uv via pip install (#681)

Changelog
- 231e622 proxy: fix matrix race and process stop bug (#677)
- 57ac666 .github/workflows: tweak push ghcr conditional (#676)
- 6972830 .github/workflows: add toggle for pushing unified images to github (#672)
- c176fa7 docker/unified: add spirv-headers to fix vulkan build (#669)

Changelog
- 5e3c646 proxy: compress captures with zstd (#668)
- c3f0d43 proxy: fix race conditions during swap (#667)
- f6cf9f5 proxy: Refactor tests (#660)
- 121fd93 Makefile: restore linux arm64 targets

Changelog
- 17233e9 docs: update configuration.md for matrix
- 4866d16 README.md: update to use matrix instead of groups
- 35193f8 proxy: add swap matrix with solver-based model swapping (#646)
- 40e39f7 ui-svelte: fix security issues (#649)

Changelog
- a9d840f proxy,proxy/config: restore timeouts to pre PR 619 (#648)
- 7b2b827 docker/unified: derive rootless image from root container (#644)

Changelog
- 8fabc75 docker/unified: vulkan build fixes (#600)
- e5e7391 .github,docker/unified: include vulkan build (#599)
- 2c282dc .github,docker/unified: improve caching and fix bugs (#598)
- 916d13f .github/workflows,docker/unified: add cuda based unified container (#597)
- a3725e7 Update go.mod to 1.26.1 (#593)
- 15bd55d proxy, ui-svelte: add /sdapi/v1 endpoint support (#587)

Changelog
- c3c258a proxy: fix metrics capture for v1/responses (#586)
- 29a38fd ui-svelte: upgrade to vite 8 (#585)
- d569681 Change model sorting to natural order (#582)
- 24efdb7 config: add macro support for name and description fields (#578)

Changelog
- cc77139 proxy,proxy/config: add global TTL feature (#554)
- 390a35b ui-svelte: add copy button to markdown code blocks (#537)
- 181f71c .github,docker: add cuda13 architecture support (#551)

Changelog
- 49546e2 ui: fix text size svg
- 2c07896 Update README with additional images
- 175bb36 Revise README description for clarity and detail
- aedb640 Enhance web UI section in README
- 2f377f6 ui: add OGG audio format support to transcription playground (#544)

Changelog
- 64e4c79 ui: add Rerank tab to playground (#536)

Changelog
- 19fb5f3 proxy: implement setParamsByID filter (#535)

Changelog
- b45102b ui: smart auto-scroll in LogPanel (#530)

Changelog
- 1688bdd proxy, ui: add pending requests count to the main dashboard (#516)

Changelog
- 4b4ee70 proxy: support timings for /infill from llama-server
- b5fde8e proxy,ui-svelte: add request/response capturing (#508)
- 7eef5de docs: add stable-diffusion.cpp references (#506)

Changelog
- bc01e6f build: add stable-diffusion server to musa and vulkan container images (#504)
- 0462e3d Reorganize UI controls and improve form interactions (#500)
- 7b20fc0 Add path filters to CI workflows and create UI test workflow (#501)

Changelog
- 20738f3 proxy,ui-svelte: replace old UI with svelte+playground
- cdea7d1 proxy/config: skip env macros in YAML comment lines (#496)

Changelog
- 5de387d ui: fix node-tar vulnerability
- 6f8e7cc .github/workflows: switch release.yml to build ui-svelte
- 4384315 ui-svelte: add Svelte port of React UI (#487)

Changelog
- 6439ab1 ui: add peer:true in package-lock.json
- f942261 build(deps-dev): bump tar from 7.5.3 to 7.5.6 in /ui (#477)
- 7493618 Add count_tokens api proxying (#476)
- 205efd4 proxy: extend /running endpoint with additional process data (#474)
- 14207f8 ui: npm security update
- 4e850c2 config: refactor macro substitution in configuration (#470)

Changelog
- 75fced5 config: support macros in peer apiKey and filters (#469)

Changelog
- 267c030 ui: update react-router-dom to 7.12.0 (#456)
- c19309f CLAUDE.md: small instruction tweaks

Changelog
- 4413881 proxy: actually add /v1/responses endpoint (#449)

Changelog
- 37d74ef proxy: add /v1/images/generations (#443)

Changelog
- 22e098a Add Peer Model Support (#438)
- 9864f9f .coderabbit.yaml: disable annoying features
- 53b32f3 proxy: add API key support (#436)

Changelog
- 565c447 config,proxy: add new configuration logToStdout (#432)
- e6a9e21 proxy: fix path bug in /logs/stream/{model_id} (#431)

Changelog
- dea9873 proxy: extract metrics for v1/messages (#419)

Changelog
- b1dec8b docker: build both root and non-root container images (#412)
- 06523d8 feat: add platform-specific process attributes support (#411)

This release includes a set of quality of life features for configuring and using llama-swap:
- add JSON schema for configuration file (#393)
- build, commit and version information in the UI (#395)
- image
- enable model aliases in v1/models (#400)
- logTimeFormat: enable and set timestamps in log output for the proxy (#401)
Shout out to @ryan-steed-usa and @nint8835 for their contributions to this release.
Changelog
- 86e9b93 proxy,ui: add version endpoint and display version info in UI (#395)
- 3acace8 proxy: add configurable logging timestamp format (#401)
- 554d29e feat: enhance model listing to include aliases (#400)
- 3567b7d Update image in README.md for web UI section
- 3873852 config.example.yaml: add modeline for schema validation
- c0fc858 Add configuration file JSON schema (#393)
- b429349 add /ui/ to wol-proxy polling (#388)
- eab2efd feat: improve llama.cpp base image tag for cpu (#391)
- 6aedbe1 cmd/wol-proxy: show a loading page for / (#381)
- b24467a fix: update containerfile user/group management commands (#379)

Changelog
- 12b69fb proxy: recover from panic in Process.statusUpdate (#378)
- f91a8b2 refactor: update Containerfile to support non-root user execution and improve security (#368)

This release includes a unique feature to show model loading progress in the Reasoning content. When enabled in the config llama-swap will stream a bit of data so there is no silence when waiting for the model to swap and load.
- Add a new global config setting: sendLoadingState: true
- Add a new model override setting: model.sendLoadingState: true to control it on per model basis
Demo:
llama-swap-issue-366.mp4
Thanks to @ServeurpersoCom for the very cool idea!
Changelog
- a89b803 Stream loading state when swapping models (#371)

Fix a bug where a panic() can cause llama-swap to lock up or exit. Recommended update.
Changelog
- f852689 proxy: add panic recovery to Process.ProxyRequest (#363)

Changelog
- e250e71 Include metrics from upstream chat requests (#361)
- d18dc26 cmd/wol-proxy: tweak logs to show what is causing wake ups (#356)

Changelog
- 8357714 ui: fix avg token/sec calculation on models page (#357)
Averages were replaced with percentiles and a histogram:
image

This release adds cmd/wol-proxy, a Wake-on-LAN proxy for llama-swap. If llama-swap lives on a high idle wattage server that suspends after an idle period, wol-proxy will automatically wake that server up and then reverse proxy the requests.
A niche use case but hopefully it will save a lot of wasted energy from idle GPUs.
Changelog
- c07179d cmd/wol-proxy: add wol-proxy (#352)
- 7ff5063 Update README for setup instructions clarity [skip ci]
- 9fc0431 Clean up and Documentation (#347) [skip ci]

This release includes support for TLS certificates from contributor @dwrz!
To use it:
./llama-swap --tls-cert-file /path/to/cert.pem --tls-key-file /path/to/key.pem ...
Generating a self-signed certificate:
openssl req -x509 -newkey rsa:4096 -keyout key.pem -out cert.pem -days 365 -nodes
Changelog
- 6516532 Add optional TLS support (#340)
- d58a8b8 Refactor to use httputil.ReverseProxy (#342)
- caf9e98 Fix race conditions in proxy.Process (#349)

Changelog
- 5392783 ui: tweak vertical space for mobile (#343)

Changelog
- 00b738c Add Macro-In-Macro Support (#337)

This release includes two new features:
- model macros (#330): macros can now be defined as part of a model's configuration. These take precedence over macros defined at the global level.
- model metadata (#333): metadata can now be defined in a model's configuration. This is a schema-less object that supports integers, floats, bools, strings, arrays and child objects. metadata fields also support macro substitution. Metadata is only used in the v1/models endpoint under a new JSON key: meta.llamaswap.
Other smaller changes:
- macro values can be any integer, string, bools, or float types. This enhancement makes JSON encoding of metadata with macros behave as expected. Previously macro values could only be strings.
Changelog
- 70930e4 proxy: add support for user defined metadata in model configs (#333)
- 1f61791 proxy/config: add model level macros (#330)
- 216c40b proxy/config: create config package and migrate configuration (#329)

Changelog
- 9e3d491 proxyToUpstream: add redirect with trailing slash to upstream endpoint (#322)