Most hosted AI chat apps trade your privacy for convenience: your conversations and keys live on someone else's server. NextChat inverts that — it ships as a thin client you deploy yourself, keeps every message in the browser's local storage, and routes requests directly to whatever provider's key you hand it. The result is a ChatGPT-style interface you fully own, deployable in one click to Vercel or Docker.
What Sets It Apart
- Provider-agnostic by design: one UI talks to OpenAI, Azure, Anthropic Claude, Google Gemini, DeepSeek, Baidu, ByteDance, Alibaba, ChatGLM and more — switch models without switching apps.
- Genuinely everywhere: the same codebase (Next.js + React, Tauri for desktop) runs as web, PWA, browser extension, and native iOS/macOS/Windows/Linux/Android builds.
- Beyond plain chat: MCP support, a plugin system, Markdown with LaTeX and code highlighting, and artifact/image-generation integration push it past a simple wrapper.
Who It's For
Great fit if you want a self-hosted, multi-provider chat front-end and prefer paying per-token through your own keys rather than a flat subscription. Look elsewhere if you need server-side conversation sync across devices, team accounts, or managed infrastructure — local-first storage means your history lives on each device, and you own the deployment and key management yourself.
