Every feature runs locally on your hardware. Choose your model engine per tool. Swap providers without changing a line of code.
These tools are production-ready and included in every install.
Upload PDFs, Word documents, spreadsheets, code files, and more. local-ai indexes your files using local embeddings, stores them in a vector database, and lets you ask natural-language questions against your data.
Convert any text into natural-sounding speech using locally-running TTS models. Multiple voices, multiple languages, zero cloud dependency. Ideal for accessibility, content creation, and audio previews.
These features are in active development. Star the repo to get notified when they ship.
Text-to-image using local Stable Diffusion, SDXL, or ComfyUI workflows. Generate, inpaint, and upscale — all on your GPU.
Drop lengthy reports, research papers, or legal docs. Get structured summaries with key takeaways, organized by section.
Index all your files and search by meaning. Surface connections across documents, find related content, and navigate your knowledge base naturally.
Community plugin marketplace. Build custom tools with the SDK and plug them into any local model. Share plugins via the registry.
local-ai doesn't lock you into one provider. Swap engines per feature without changing your workflow.
Easiest setup. Pull models with one command. Great for getting started.
Full SupportGUI-based model manager with OpenAI-compatible server built in.
Full SupportHigh-throughput serving with PagedAttention. Best for multi-user setups.
Full SupportLightweight C++ inference. Runs on CPU, Apple Silicon, and CUDA.
Full SupportAny server exposing the /v1/chat/completions endpoint works out of the box.
Full SupportWrite a thin adapter using our Engine SDK. Connect any inference backend.
Coming SoonGet up and running in under 2 minutes.