Douyin_TikTok_Download_API: Self-Hosted Scraper & MCP Server for No-Watermark Video
Douyin_TikTok_Download_API: Self-Hosted Scraper & MCP Server for No-Watermark Video
Douyin_TikTok_Download_API is a Python-based, self-hosted solution for scraping and downloading content from Douyin and TikTok. It offers an async REST API, a CLI, and a web console, designed for developers who need to collect and archive video and image albums without watermarks directly to their own storage. The project, with over 20,000 GitHub stars, provides a one-command docker compose up deployment, including a self-healing identity pool and PostgreSQL archive.
Connecting via MCP
The Douyin_TikTok_Download_API exposes its core service layer through an MCP endpoint at /mcp. This allows MCP clients to interact directly with the scraper's capabilities, sharing the same underlying logic as the project's REST API. Developers can find client setup guidance within the console at /mcp-guide.
The MCP integration provides access to the main capabilities:
- Parsing: Process links, share text, short links, or bare post IDs.
- Archiving: Store all parsed content, ensuring data persistence even if upstream posts are deleted.
- Downloading: Save media to local disk, with bulk download by author, skipping existing files, and duplicate cleanup.
- Watching: Monitor specific authors or posts and re-collect content on a timer.
Available Interfaces and Tools
Beyond MCP, Douyin_TikTok_Download_API offers several other entry points for interaction:
- Web console: Accessible at
/, this interface provides access to the identity pool, scheduler, content library, downloads, logs, and diagnostics. - API reference: Swagger UI is available within the console at
/docs, with bare reference at/swaggerand/redocrequiring no login. - REST API: A comprehensive set of 93 operations is exposed under
/api/v1/.... - CLI: The command-line interface, invoked with
dtk --help, mirrors the functionality of the other entry points. - iOS Shortcut support: Specific integration is available via
/api/v1/ios/shortcut.
Core Capabilities for Media Management
This MCP Server focuses on robust, self-managed media collection. It fetches posts, authors, comments, and search results. A key feature is its ability to download videos and image albums without watermarks by selecting the clean stream already published by the platform, rather than attempting to strip watermarks. All collected data is stored in your own PostgreSQL database, ensuring control and ownership. The system is designed for self-hosting, eliminating external quotas or sign-ups.