LocalAI

LocalAI is an open source AI engine, released under the MIT license, that runs models on the user's own hardware through one runtime. It handles language models, vision, speech recognition and synthesis, image, video, music and 3D generation, and agent workloads, and it is designed to work on a laptop with no GPU while scaling up to GPU clusters. The core is a single binary with an OpenAI compatible API, and it also speaks the Anthropic, Ollama and ElevenLabs APIs. Existing clients usually need only a changed URL. The engine behind each model is swappable through one line of model configuration, with backends including llama.cpp, vLLM, SGLang and MLX, and backends are pulled in only when a model needs them, which keeps the base install small. Realtime voice over WebRTC, transcription with diarization, speech synthesis with voice cloning, tool calling, structured output, MCP, agents and RAG are covered, and a gallery lists more than a thousand models that install in one click. Supported hardware paths include x86_64, ARM64, CUDA, ROCm, SYCL, Metal and Vulkan. LocalAI suits developers and organizations that want private, self hosted inference as a drop in replacement for hosted model APIs. The software is free to download and use. The site publishes no pricing page, and invites businesses to get in touch for commercial discussions.