OSS Tanbou
View on GitHub

LocalAI — A self-hosted AI engine that composes LLM, vision, voice, image, and video backends behind familiar APIs

OSS health 96
About these scores

Discovery score is an unbounded weighted, log-compressed index of stars, watchers, forks, and contributors. Growth momentum is its change over the observed period; OSS health is a 0–100 score from available repository recency, Community Health, and release data.

Stars
49,141
Primary language
Go
License
MIT
Repository last updated
Sep 17, 2026

Overview

LocalAI is a Go-based AI engine that pulls specialized backends such as llama.cpp, vLLM, whisper.cpp, Stable Diffusion, and MLX only when a model needs them. It presents LLM, vision, voice, image, and video capabilities through OpenAI-, Anthropic-, and ElevenLabs-compatible APIs. The latest release checked on September 17, 2026 is v4.9.0.

Features and best fit

Based on official documentation; not hands-on tested · Content checked:

Pull only the backend a model needs and use one API across modalities and hardware

Rather than shipping one huge bundle, LocalAI isolates engines such as llama.cpp, vLLM, whisper.cpp, Stable Diffusion, and MLX into on-demand backends. It supports CPU-only as well as NVIDIA, AMD, Intel, Apple Silicon, and Vulkan paths.

Sources: [2]

Add agents, RAG, MCP, authentication, quotas, and roles above compatible inference APIs

Compatibility APIs make existing clients easier to reuse, while LocalAI also provides agents with tools, RAG, MCP, API-key authentication, usage quotas, and role-based access.

Sources: [2]

For teams that want one self-hosted control layer across multiple local AI runtimes and modalities

It fits environments that want text, speech, vision, and media generation under common APIs and access control while selecting specialized runtimes per model.

Sources: [2]

Resource and compatibility requirements remain backend- and model-specific

Container images, GPU drivers, RAM/VRAM, and model licenses still need to be evaluated per backend. Enable authentication and quotas before exposing a service remotely, and pin backend/model versions when reliability matters because releases evolve quickly.

Sources: [2][3]

Official sources

  1. [1]mudler/LocalAI — GitHub repository metadata(2026-09-17)
  2. [2]LocalAI — README(2026-09-17)
  3. [3]LocalAI v4.9.0 release(2026-09-17)
Supplemental curator note

The key distinction is that LocalAI is a control layer across multiple specialized AI backends, not a single model runtime.

Try it in 3 steps

  1. 1

    Get the source

    git clone --depth 1 https://github.com/mudler/LocalAI.git
  2. 2

    Enter the repository

    cd LocalAI
  3. 3

    Check the official steps

    Continue with the commands in the README Installation, Quick Start, or Getting Started section.

    find . -maxdepth 1 -iname 'README*' -exec sed -n '1,220p' {} \; -quit
Check the official README

Growth

Growth trends · Last 30 days

49,141 Stars

Trend data is still being collected.

Development activity

Last 90 days · weekly

Commits (last 30 days)
412
Open PRs
70

Development activity is still being collected.

Built with

Categories and tags

Categories

Tags

GitHub data

GitHub dataView detailed GitHub data
View on GitHub

GitHub Topics

  • llama
  • ai
  • llm
  • stable-diffusion
  • api
  • tts
  • musicgen
  • mamba
  • audio-generation
  • image-generation
  • text-generation
  • rerank
Stars
49,141
Forks
4,455
Watchers
300
Open issues
98
Primary language
Go
License
MIT
Repository last updated
Sep 17, 2026
Write a related article

Share a guide or use case for this OSS in Markdown. Articles are published after administrator approval.

Report incorrect information

Tell us if any listing information is incorrect or outdated.