YanJiuWeiYuanHui/DeepChat

DeepChat extends DeepSeek's static web chat project, supporting DeepSeek's native interface, OpenAI Chat Completions and Responses interfaces, as well as key-free local service presets with automatic model discovery. Speech recognition, speech synthesis, image recognition, OCR, and image generation can be configured separately; it also features continuous voice calls, long-term memory, per-turn token usage tracking, and daily usage alerts. Chat history can be imported and exported in DeepSeek's official archive format.

DeepChat extends DeepSeek's static web chat project, supporting DeepSeek's native interface, OpenAI Chat Completions and Responses interfaces, as well as key-free local service presets with automatic model discovery. Speech recognition, speech synthesis, image recognition, OCR, and image generation can be configured separately; it also features continuous voice calls, long-term memory, per-turn token usage tracking, and daily usage alerts. Chat history can be imported and exported in DeepSeek's official archive format.

Why it matters In addition to DeepSeek, it supports OpenAI-compatible interfaces and local services such as Ollama and LM Studio, allowing speech, image recognition, and image generation to be configured with different providers.

Stars 25

Reasons

  • MIT license; Releases provide pre-compiled Android APKs
  • The README provides steps for local development, static builds, and automated signed releases

Usage

  • Requires Node.js 22 or higher and pnpm: after git clone, run pnpm install and pnpm dev, open localhost:8090 and enter your API Key; for mobile devices, download the APK from Releases

Risks

  • Maintained by an individual developer; the author welcomes bug reports; API Keys are stored on the client side, so do not save them on public devices
  • When self-hosting, do not expose the built-in /__api_proxy__ forwarding endpoint to the public internet

Open

GitHub

Explore the latest AI Radar updates