docsivx/ai Chat

Documentation

Installing: Install ivx/ai Chat on macOS, Windows, Linux, Android and iOS, or run it in Chrome, Firefox and Safari.
Ollama: Run open-weight models on your own machine or on Ollama Cloud, and point ivx/ai Chat at them.; not needed / required for Cloud
LM Studio: Point ivx/ai Chat at a model served locally by LM Studio.; not needed
llama.cpp: Point ivx/ai Chat at a llama.cpp server running on your machine.; not needed
OpenRouter: Reach many hosted models through a single OpenRouter key.; openrouter.ai/keys
OpenAI: Use an OpenAI API key with ivx/ai Chat.; required
Anthropic: Use an Anthropic API key with ivx/ai Chat.; required
Groq, Mistral, Together and DeepSeek: Add further hosted providers from the preset list, or any OpenAI-compatible endpoint.; required
The CORS bridge: A small daemon that lets the browser reach endpoints it would otherwise refuse to call — Ollama on its defaults, a bare llama.cpp build, an old company proxy.
Self-hosting: Run your own copy of ivx/ai Chat: static files on any host, or the bridge serving the app and the proxy together, on a VPS or your own computer.
Building it: Build the web app, the CORS bridge and the desktop and mobile apps from source, and how a release is cut.