anYuJia/better-douyin

better‑douyin – a lightweight Rust/Tauri desktop client for Douyin (TikTok)

What it is

• A desktop application (Windows/macOS/Linux) built with Rust + Tauri for the runtime and React + Vite + TypeScript for the UI.
• It provides a full‑featured Douyin experience: browsing feeds, searching users, watching videos, downloading content, managing multiple accounts, and even AI‑assisted interactions (e.g., auto‑generated comments or replies via an OpenAI‑compatible service).
• The public repository only contains the "open shell" – the UI, mock bridge, and a mock Node.js backend that returns demo data. The real, functional backend (which talks to Douyin’s servers, handles cookies, signatures, etc.) is shipped as a compiled binary in the GitHub Releases.

Why it exists

• To give users a lighter, more stable desktop experience than the official mobile/web client, with native playback, offline download management, and cross‑platform distribution.
• To keep the core logic closed (to avoid abuse) while still allowing the community to contribute UI/UX improvements, theme tweaks, accessibility fixes, and other front‑end enhancements.

Key capabilities (in the full binary)

• Content browsing – search by username/UID, parse share links, view user profiles, feed, liked, favorited, and collection videos.
• Download manager – queue, pause/resume, batch download, custom naming templates, quality selection, and local library view.
• Embedded player – speed, volume, resolution switch, auto‑play next, and in‑player actions (like, comment, share).
• Social features – friends list, private messages, notifications, and AI‑driven comment/DM generation.
• Automation – monitor feeds, private messages, or creator updates; trigger auto‑like, auto‑comment, auto‑follow, etc., based on user‑defined rules.
• Multi‑account support – store cookies and account data locally, switch accounts safely.
• AI integration – plug in any OpenAI‑compatible service to generate text for comments, replies, or content analysis.
• Local HTTP MCP service – lets external AI tools (e.g., Claude, Codex) query the app’s state or issue commands with explicit user consent.

How to try it

1. Download the compiled release from the Releases page (the repository’s README strongly recommends this for ordinary users). The release includes checksum files for verification.
2. Run the installer; no building from source is required.
3. For developers, you can run the UI demo locally:

   npm --prefix frontend install
   npm run dev   # starts the React dev server with a mock backend
   

   The mock backend (backend/server.mjs) only serves static demo data and does not contact Douyin.

Open‑source scope

• The repo contains:
  • Front‑end code (React components, state management, styling).
  • A mock Node.js server that mimics the API contract.
  • Documentation describing the public‑vs‑private boundaries.
• It does not include:
  • Real Douyin API calls, authentication cookies, signatures, or decryption logic.
  • Any code that would violate Douyin’s terms of service (e.g., mass scraping, anti‑rate‑limit bypasses).

License

• Distributed under a Better Douyin Non‑Commercial License – free for personal learning, research, or testing, but commercial redistribution or SaaS hosting is prohibited.

Who might benefit

• Power users who want a desktop‑first Douyin client with download management.
• Developers interested in contributing UI/UX improvements or building their own backend that adheres to the public API contract.
• AI enthusiasts who want to experiment with AI‑generated comments or integrate the app with other AI tools via the provided local HTTP interface.

---

All details above are taken directly from the repository’s README; no additional features have been inferred.