NextChat
A self-hostable AI chat interface with Docker deployment, local model connections through LocalAI and RWKV-Runner, and desktop clients.
NextChat is a self-hosted AI chat interface for people who want one place to use their own LLM server and cloud models. The web and desktop project is open source under the MIT license. You can host it with Docker or on Vercel, and desktop clients run on macOS, Windows and Linux.
Local model connections work with LocalAI and RWKV-Runner. For cloud inference, it supports providers including OpenAI, Anthropic, Google Gemini and DeepSeek, plus Azure endpoints. The interface stores chat data locally in the browser, but cloud model requests still go to the selected provider. Self-hosting the interface doesn't make those requests local.
Reusable prompt templates let you keep instructions for recurring tasks and share them with others. NextChat streams responses as they arrive and automatically compresses conversation history to keep long chats within the model's context. It renders Markdown with LaTeX equations, Mermaid diagrams and syntax highlighting, so technical answers can retain their formatting.
MCP support and plugins connect chat to tools such as web search, calculators and other APIs. Artifacts give generated content and webpages a separate preview window where you can copy or share them. The interface includes multiple languages, dark mode and a responsive browser layout with PWA support. You can also share conversations as images or through ShareGPT.
