HiveChat

Why we included this project

HiveChat is a self-hosted chat app for small and medium teams that want one shared AI workspace without each member managing their own API keys. A single administrator configures the model providers once, then assigns users to groups with per-group model visibility and monthly token limits, which keeps cost and access under control. It connects to a wide range of providers, including OpenAI, Claude, Gemini, DeepSeek, and many Chinese services, and accepts any OpenAI-style endpoint, so teams can mix vendors freely. Beyond plain chat it supports MCP servers over SSE, image understanding, LaTeX and Markdown rendering, and DeepSeek chain-of-thought display, with email or WeCom, DingTalk, and Feishu login options. For teams that want a central admin to control which models people can use and how much they can spend, rather than handing out API keys to everyone, this is a practical option.