2 Releases published by 1 person

• v3.3.1

  published Aug 1, 2025

• v3.3.2

  published Aug 4, 2025

27 Pull requests merged by 4 people

• chore: ⬆️ Update ggml-org/llama.cpp to 41613437ffee0dbccad684fc744788bc504ec213

  #5968 merged Aug 4, 2025

• docs: ⬆️ update docs version mudler/LocalAI

  #5967 merged Aug 4, 2025

• chore(build): Rename sycl to intel

  #5964 merged Aug 4, 2025

• chore: ⬆️ Update ggml-org/llama.cpp to d31192b4ee1441bbbecd3cbf9e02633368bdc4f5

  #5965 merged Aug 3, 2025

• feat(backends): allow backends to not have a metadata file

  #5963 merged Aug 3, 2025

• feat(backends): install from local path

  #5962 merged Aug 3, 2025

• chore(stable-diffusion): bump, set GGML_MAX_NAME

  #5961 merged Aug 3, 2025

• chore: ⬆️ Update ggml-org/llama.cpp to 5c0eb5ef544aeefd81c303e03208f768e158d93c

  #5959 merged Aug 2, 2025

• chore: ⬆️ Update ggml-org/whisper.cpp to 0becabc8d68d9ffa6ddfba5240e38cd7a2642046

  #5958 merged Aug 2, 2025

• docs: ⬆️ update docs version mudler/LocalAI

  #5956 merged Aug 1, 2025

• fix(docs): Improve responsiveness of tables

  #5954 merged Aug 1, 2025

• chore: ⬆️ Update ggml-org/llama.cpp to daf2dd788066b8b239cb7f68210e090c2124c199

  #5951 merged Aug 1, 2025

• feat(swagger): update swagger

  #5950 merged Jul 31, 2025

• fix(intel): Set GPU vendor on Intel images and cleanup

  #5945 merged Jul 31, 2025

• chore(model gallery): add flux.1-krea-dev-ggml

  #5949 merged Jul 31, 2025

• chore(model gallery): add flux.1-dev-ggml-abliterated-v2-q8_0

  #5948 merged Jul 31, 2025

• chore(model gallery): add flux.1-dev-ggml-q8_0

  #5947 merged Jul 31, 2025

• chore(capability): improve messages

  #5944 merged Jul 31, 2025

• feat(stablediffusion-ggml): allow to load loras

  #5943 merged Jul 31, 2025

• chore: update swagger

  #5946 merged Jul 31, 2025

• chore: ⬆️ Update ggml-org/whisper.cpp to f7502dca872866a310fe69d30b163fa87d256319

  #5941 merged Jul 31, 2025

• chore: ⬆️ Update ggml-org/llama.cpp to e9192bec564780bd4313ad6524d20a0ab92797db

  #5940 merged Jul 31, 2025

• feat(stablediffusion-ggml): add support to ref images (flux Kontext)

  #5935 merged Jul 30, 2025

• chore(model gallery): add qwen_qwen3-30b-a3b-thinking-2507

  #5939 merged Jul 30, 2025

• chore(model gallery): add arcee-ai_afm-4.5b

  #5938 merged Jul 30, 2025

• chore(model gallery): add qwen_qwen3-30b-a3b-instruct-2507

  #5936 merged Jul 30, 2025

• chore: ⬆️ Update ggml-org/llama.cpp to aa79524c51fb014f8df17069d31d7c44b9ea6cb8

  #5934 merged Jul 29, 2025

3 Pull requests opened by 1 person

• chore(deps): bump torch and sentence-transformers

  #5969 opened Aug 4, 2025

• chore(deps): bump torch and diffusers

  #5970 opened Aug 4, 2025

• feat(backends install): allow to specify name and alias during manual installation

  #5971 opened Aug 4, 2025

13 Issues closed by 2 people

• error="unexpected end of JSON input"

  #2850 closed Aug 5, 2025

• Unable to find image 'localai/localai:21' locally

  #2864 closed Aug 5, 2025

• diffuser backend processes stack up and hog GPU memory

  #2866 closed Aug 5, 2025

• Document storage requirement

  #2869 closed Aug 4, 2025

• Document permission requirements

  #2870 closed Aug 4, 2025

• Allow sideloading backends on filesystem

  #5917 closed Aug 3, 2025

• unable to load hercules-5.0-qwen2-7b with docker

  #2900 closed Aug 3, 2025

• Must we need docker file to run LocalAI?

  #2913 closed Aug 3, 2025

• Build error in build with GO_TAGS=stablediffusion

  #2934 closed Aug 3, 2025

• intfloat/multilingual-e5-base can't load

  #2946 closed Aug 3, 2025

• Fails: could not load model: rpc error: code = Unavailable desc = error reading from server: EOF

  #3007 closed Aug 3, 2025

• Error generating images with LocalAI integrated with Nextcloud AI (CPU only)

  #3017 closed Jul 30, 2025

• Misspelling of the word "adequately"

  #3028 closed Jul 30, 2025

3 Issues opened by 3 people

• gRPC update to 1.74 breaks local-ai on metal

  #5966 opened Aug 4, 2025

• whisper backend not available on apple silicon (m4 pro)

  #5953 opened Aug 1, 2025

• CPU-only version on Kubernetes references nvidia - and fails

  #5952 opened Aug 1, 2025

11 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• old torch causes version restriction due to CVE

  #5933 commented on Jul 29, 2025 • 0 new comments

• Unable to generate sd3 medium images with 24gb gpu

  #2845 commented on Aug 1, 2025 • 0 new comments

• animagine-xl wont start in "aio-gpu-nvidia-cuda-12"

  #2787 commented on Aug 2, 2025 • 0 new comments

• "Error: Failed to process stream" appears when on the "Chat" tab trying to chat to any model.

  #3001 commented on Aug 3, 2025 • 0 new comments

• Invalid json crashes the server / Json schema suppport incomplete

  #2938 commented on Aug 3, 2025 • 0 new comments

• ERR error installing backend - failed to download layer 0: unexpected EOF

  #5924 commented on Aug 3, 2025 • 0 new comments

• The prompt execution is not interrupted where the conversation is cleared in the WebUI

  #2731 commented on Aug 4, 2025 • 0 new comments

• Create authorization bearer tokens directly in the WebUI

  #2730 commented on Aug 4, 2025 • 0 new comments

• There is something wrong with VLM

  #2668 commented on Aug 5, 2025 • 0 new comments

• Can't build LocalAI with llama.cpp with CUDA

  #3418 commented on Aug 5, 2025 • 0 new comments

• ci: add static-checker

  #2781 commented on Aug 4, 2025 • 0 new comments