Skip to content

LocalAI

v4.3.2 Feature

This release adds 5 notable features for engineering teams evaluating rollout.

Published 7d Model Serving & MLOps
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

agents ai api audio-generation decentralized distributed
+12 more
image-generation libp2p llama llm mamba mcp musicgen object-detection rerank stable-diffusion text-generation tts

Affected surfaces

deps

Summary

AI summary

Broad release touches Other Changes, chore, 👒 Dependencies, and deps.

Changes in this release

Feature Medium

Add Model routing, PII filtering, and Cloud model proxies middleware.

Add Model routing, PII filtering, and Cloud model proxies middleware.

Source: llm_adapter@2026-05-27

Confidence: high

 —
Feature Medium

Introduce gated X‑LocalAI‑Node response header (middleware + wrapper).

Introduce gated X‑LocalAI‑Node response header (middleware + wrapper).

Source: llm_adapter@2026-05-27

Confidence: high

 —
Feature Medium

Add LTX‑2 support and LTX‑2.3 GGUF gallery entries for stablediffusion‑ggml.

Add LTX‑2 support and LTX‑2.3 GGUF gallery entries for stablediffusion‑ggml.

Source: llm_adapter@2026-05-27

Confidence: high

 —
Feature Low

Mux LTX‑2 audio into the output MP4 for stablediffusion‑ggml.

Mux LTX‑2 audio into the output MP4 for stablediffusion‑ggml.

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Feature Low

Update Swagger documentation.

Update Swagger documentation.

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Feature Low

Add native object detection and segmentation backend (rfdetr‑cpp).

Add native object detection and segmentation backend (rfdetr‑cpp).

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Dependency Low

Bump github.com/nats-io/nats.go from 1.50.0 to 1.52.0.

Bump github.com/nats-io/nats.go from 1.50.0 to 1.52.0.

Source: llm_adapter@2026-05-27

Confidence: high

 —
Dependency Low

Bump github.com/aws/aws-sdk-go-v2/credentials from 1.19.15 to 1.19.17.

Bump github.com/aws/aws-sdk-go-v2/credentials from 1.19.15 to 1.19.17.

Source: llm_adapter@2026-05-27

Confidence: high

 —
Dependency Low

Bump actions/stale from 10.2.0 to 10.3.0.

Bump actions/stale from 10.2.0 to 10.3.0.

Source: llm_adapter@2026-05-27

Confidence: high

 —
Dependency Low

Bump sentence-transformers from 5.5.0 to 5.5.1 in /backend/python/transformers.

Bump sentence-transformers from 5.5.0 to 5.5.1 in /backend/python/transformers.

Source: llm_adapter@2026-05-27

Confidence: high

 —
Dependency Low

Update transformers requirement from >=5.8.1 to >=5.9.0 in /backend/python/transformers.

Update transformers requirement from >=5.8.1 to >=5.9.0 in /backend/python/transformers.

Source: llm_adapter@2026-05-27

Confidence: high

 —
Dependency Low

Bump protobuf from 6.33.5 to 7.35.0 in /backend/python/transformers.

Bump protobuf from 6.33.5 to 7.35.0 in /backend/python/transformers.

Source: llm_adapter@2026-05-27

Confidence: high

 —
Bugfix Medium

Fix VRAM detection in intel component.

Fix VRAM detection in intel component.

Source: llm_adapter@2026-05-27

Confidence: high

 —
Bugfix Low

Persist per-model load info so distributed reconciler survives frontend restart.

Persist per-model load info so distributed reconciler survives frontend restart.

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Bugfix Low

Sync gallery OpCache and caches across frontend replicas in distributed mode.

Sync gallery OpCache and caches across frontend replicas in distributed mode.

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Bugfix Low

Add diffusion_model flag to all LTX‑2.3 variants in the model gallery.

Add diffusion_model flag to all LTX‑2.3 variants in the model gallery.

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Bugfix Low

Set vae_decode_only:false for i2v and flf2v variants of LTX‑2.3.

Set vae_decode_only:false for i2v and flf2v variants of LTX‑2.3.

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Bugfix Low

Prevent thinkstep leakage into content when autoparser is in pure-content mode (reasoning).

Prevent thinkstep leakage into content when autoparser is in pure-content mode (reasoning).

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Bugfix Low

Exclude local‑only artifacts from Docker build context via .dockerignore.

Exclude local‑only artifacts from Docker build context via .dockerignore.

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Bugfix Low

Share a single /api/operations poller across React‑UI consumers to reduce load.

Share a single /api/operations poller across React‑UI consumers to reduce load.

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Bugfix Low

Polish 'Fits in my GPU' filter to use design‑system Toggle component.

Polish 'Fits in my GPU' filter to use design‑system Toggle component.

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Bugfix Low

Force .check() on hidden Toggle input for e2e testing of the fits‑filter feature.

Force .check() on hidden Toggle input for e2e testing of the fits‑filter feature.

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Other Low

Add 'Fits in my GPU' filter on Install Models UI page.

Add 'Fits in my GPU' filter on Install Models UI page.

Source: granite4.1:30b@2026-05-27-audit

Confidence: low

 —
Full changelog

What's Changed

👒 Dependencies

  • chore(deps): bump github.com/nats-io/nats.go from 1.50.0 to 1.52.0 by @dependabot[bot] in https://github.com/mudler/LocalAI/pull/10003
  • chore(deps): bump github.com/aws/aws-sdk-go-v2/credentials from 1.19.15 to 1.19.17 by @dependabot[bot] in https://github.com/mudler/LocalAI/pull/10008
  • chore(deps): bump actions/stale from 10.2.0 to 10.3.0 by @dependabot[bot] in https://github.com/mudler/LocalAI/pull/10002
  • chore(deps): bump sentence-transformers from 5.5.0 to 5.5.1 in /backend/python/transformers by @dependabot[bot] in https://github.com/mudler/LocalAI/pull/10007
  • chore(deps): update transformers requirement from >=5.8.1 to >=5.9.0 in /backend/python/transformers by @dependabot[bot] in https://github.com/mudler/LocalAI/pull/10005
  • chore(deps): bump protobuf from 6.33.5 to 7.35.0 in /backend/python/transformers by @dependabot[bot] in https://github.com/mudler/LocalAI/pull/10004

Other Changes

  • feat(middleware): Model routing, PII filtering, Cloud model proxies by @richiejp in https://github.com/mudler/LocalAI/pull/9802
  • fix(intel): VRAM detection by @richiejp in https://github.com/mudler/LocalAI/pull/9944
  • feat(distributed): gated X-LocalAI-Node response header (middleware + wrapper) by @localai-bot in https://github.com/mudler/LocalAI/pull/9976
  • fix(distributed): persist per-model load info so reconciler survives frontend restart by @localai-bot in https://github.com/mudler/LocalAI/pull/9981
  • feat(stablediffusion-ggml): LTX-2 support + LTX-2.3 GGUF gallery entries by @localai-bot in https://github.com/mudler/LocalAI/pull/9980
  • fix(distributed): sync gallery OpCache + caches across frontend replicas by @localai-bot in https://github.com/mudler/LocalAI/pull/9983
  • fix(gallery/ltx-2.3): add diffusion_model flag to all variants by @mudler in https://github.com/mudler/LocalAI/pull/9986
  • fix(gallery/ltx-2.3): add vae_decode_only:false for i2v / flf2v by @mudler in https://github.com/mudler/LocalAI/pull/9987
  • fix(reasoning): stop leaking into content when autoparser is in pure-content mode by @localai-bot in https://github.com/mudler/LocalAI/pull/9991
  • fix(stablediffusion-ggml): mux LTX-2 audio into output MP4 by @localai-bot in https://github.com/mudler/LocalAI/pull/9990
  • feat(swagger): update swagger by @localai-bot in https://github.com/mudler/LocalAI/pull/9992
  • docs: :arrow_up: update docs version mudler/LocalAI by @localai-bot in https://github.com/mudler/LocalAI/pull/9993
  • fix(streaming/tools): stop healing-marker stubs from gating off content by @localai-bot in https://github.com/mudler/LocalAI/pull/9999
  • chore: :arrow_up: Update antirez/ds4 to ad0209f6a4b067574d2b4afe896c08c177156b31 by @localai-bot in https://github.com/mudler/LocalAI/pull/9996
  • chore: :arrow_up: Update ikawrakow/ik_llama.cpp to b4e1d916c5ec7e75ea3c124dd090425a99fc613f by @localai-bot in https://github.com/mudler/LocalAI/pull/9995
  • chore: :arrow_up: Update leejet/stable-diffusion.cpp to 1ceb5bd9df7784bcdf67dd9ed8bf0198b542ebc9 by @localai-bot in https://github.com/mudler/LocalAI/pull/9994
  • chore: :arrow_up: Update ggml-org/whisper.cpp to e0fd1f6787a5bd4a4957dd97c5b64df882ee7b0c by @localai-bot in https://github.com/mudler/LocalAI/pull/9997
  • fix(streaming/tools): don't leak prefill-misclassified content as trailing reasoning chunk by @localai-bot in https://github.com/mudler/LocalAI/pull/10000
  • chore: :arrow_up: Update ggml-org/llama.cpp to 35c9b1f39ebe5a7bb83986d64415a079218be78d by @localai-bot in https://github.com/mudler/LocalAI/pull/9998
  • chore(model gallery): :robot: add 1 new models via gallery agent by @localai-bot in https://github.com/mudler/LocalAI/pull/10011
  • fix(dockerignore): exclude local-only artifacts from build context by @richiejp in https://github.com/mudler/LocalAI/pull/10015
  • chore(model gallery): :robot: add 1 new models via gallery agent by @localai-bot in https://github.com/mudler/LocalAI/pull/10016
  • test: add Go + React UI coverage gates and fill test gaps by @richiejp in https://github.com/mudler/LocalAI/pull/9989
  • fix(qwen-asr): enable timestamp output when forced_aligner is configured by @fqscfqj in https://github.com/mudler/LocalAI/pull/10013
  • fix(nemo): extract Hypothesis.text for TDT/RNNT ASR models by @fqscfqj in https://github.com/mudler/LocalAI/pull/10012
  • chore: :arrow_up: Update ikawrakow/ik_llama.cpp to d2da6da05c73aeb658a3d1751f386c24e6963856 by @localai-bot in https://github.com/mudler/LocalAI/pull/10020
  • chore: :arrow_up: Update ggml-org/whisper.cpp to 27101c01dcac1676e2b6422256233cd0f1f9ae28 by @localai-bot in https://github.com/mudler/LocalAI/pull/10021
  • chore: :arrow_up: Update ggml-org/llama.cpp to 0d18aaa9d1a8af3df9abccd828e22eeaac7f840b by @localai-bot in https://github.com/mudler/LocalAI/pull/10022
  • chore: :arrow_up: Update leejet/stable-diffusion.cpp to 92dc7268fc4ffb0c0cc0bd52dfcefea91326e797 by @localai-bot in https://github.com/mudler/LocalAI/pull/10023
  • chore: :arrow_up: Update antirez/ds4 to e8e8779b261c10f36ad6270ba732c8f0be5b62e3 by @localai-bot in https://github.com/mudler/LocalAI/pull/10024
  • UI: add 'Fits in my GPU' filter on Install Models by @siddimore in https://github.com/mudler/LocalAI/pull/10017
  • fix(react-ui): share single /api/operations poller across consumers by @localai-bot in https://github.com/mudler/LocalAI/pull/10029
  • feat(backend): rfdetr-cpp native object detection + segmentation backend by @localai-bot in https://github.com/mudler/LocalAI/pull/10028
  • fix(react-ui): polish 'Fits in my GPU' filter to use design-system Toggle by @localai-bot in https://github.com/mudler/LocalAI/pull/10030
  • fix(react-ui): force .check() on hidden Toggle input in fits-filter e2e by @localai-bot in https://github.com/mudler/LocalAI/pull/10031

New Contributors

  • @fqscfqj made their first contribution in https://github.com/mudler/LocalAI/pull/10013

Full Changelog: https://github.com/mudler/LocalAI/compare/v4.3.1...v4.3.2

View diff on GitHub

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Share on X Share on Bluesky

Track LocalAI

Get notified when new releases ship.

Sign up free

About LocalAI

LocalAI is the open-source AI engine. Run any model - LLMs, vision, voice, image, video - on any hardware. No GPU required.

All releases →

Related context

Beta — feedback welcome: [email protected]