LocalAI Download: A Self-Hosted, OpenAI-Compatible AI Server
LocalAI · for Linux, macOS and Windows
Developer: Ettore Di Giacinto and the LocalAI team (opens in a new tab)
- Linux
- macOS
- Windows
The download link points to the program’s official source; the file is not hosted on this site. Link address: github.com
Link not working? Let us know
LocalAI is a free, open-source engine that runs language, image, audio and video models on your own hardware and exposes them through OpenAI-, Anthropic- and ElevenLabs-compatible APIs. It runs via Docker or a macOS app and is MIT-licensed.
Many apps are built to talk to the OpenAI API. LocalAI lets you point them at your own machine instead: it speaks the same API but runs open models locally. It was created by Ettore Di Giacinto and is maintained by the LocalAI team.
What does LocalAI do?
- Drop-in API: OpenAI-, Anthropic- and ElevenLabs-compatible endpoints, so existing apps and libraries work with little or no change.
- Many kinds of models: LLMs, vision, voice, image and video models behind one API.
- Backends on demand: A small core pulls in backends such as llama.cpp, vLLM and whisper.cpp only when a model needs them, and detects your GPU to choose the right one.
- Agents and tools: Built-in agents with tool use, RAG and skills, plus Model Context Protocol (MCP) support.
- Multi-user: API key authentication, user quotas and role-based access.
Who uses LocalAI?
It is aimed at developers and self-hosters who want one server for several AI tasks, running on their own PC or server instead of a paid cloud API.
Starting LocalAI with Docker
The quickest route is Docker. CPU only:
docker run -ti --name local-ai -p 8080:8080 localai/localai:latest
With an NVIDIA graphics card (CUDA 12):
docker run -ti --name local-ai -p 8080:8080 --gpus all localai/localai:latest-gpu-nvidia-cuda-12
The README lists separate images for CUDA 13, AMD (ROCm), Intel (oneAPI), Vulkan and NVIDIA Jetson. Then open http://localhost:8080. On macOS there is also a DMG app; it is not signed by Apple, so after installing it you have to remove the quarantine flag as described in the README.
Hardware support
- GPU: Not required. A CPU-only image is provided.
- Acceleration: NVIDIA (CUDA 12/13), AMD (ROCm), Intel (oneAPI/SYCL), Apple Silicon (Metal), Vulkan and NVIDIA Jetson are supported.
- Memory and disk: No fixed minimum is published; they depend on the models you load.
LocalAI is free and open source under the MIT license.
Pros
- Works as a drop-in for OpenAI-compatible apps
- Text, image and audio models in one server
- Runs on CPU alone; many GPU types supported
Cons
- Docker knowledge helps for setup
- The macOS app is not signed by Apple
Frequently Asked Questions
Can I use LocalAI without a graphics card?
Yes. The project provides a CPU-only Docker image and states that no GPU is required, although a supported GPU makes responses much faster.
Will my OpenAI-based app work with LocalAI?
In most cases, yes. LocalAI exposes an OpenAI-compatible API, so you usually only change the API base address to your LocalAI server, for example http://localhost:8080.
LocalAIOpen source
Download (opens in a new tab)
