Self-hosted TikTok, Douyin and Bilibili scraper with a self-healing identity pool and MCP server.
Asynchronous data scraper featuring a self-healing identity pool, PostgreSQL archive, and MCP integration.
This tool is a self-hosted API and data scraper designed to pull posts, comments, live rooms, and no-watermark videos from TikTok, Douyin, and Bilibili. It suits developers and technical professionals in AI and gaming who need automated access to social media video data without vendor rate-limit walls.
It runs via Docker Compose, utilizing a headless browser to mint guest identities that automatically replenish when health drops. It parses signatures in pure Python (including X-Bogus, A_Bogus, and Bilibili's WBI), archives everything to PostgreSQL, and offers offline speech-to-text conversion for downloaded video assets.
It suits developers building AI agents or data pipelines requiring clean social video inputs. It does not suit non-technical users looking for a plug-and-play desktop downloader or anyone unwilling to maintain containerized infrastructure.
Headless browsers mint guest identities and top up the pool automatically when usable tokens run low.
Built-in MCP server allows AI agents to directly drive the scraping workflow and query data.
Implements a_bogus, X-Bogus, X-Gnarly, X-Dynosaur, and WBI signatures natively in Python.
Persists parsed posts, comments, and media so data remains accessible even if deleted upstream.
Free and open source under its stated licence, with the only cost being the server resources and database infrastructure you provision to run it.
879 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
A headless browser automatically mints guest identities and replenishes the pool when health tiers drop or endpoints rate-limit.
Yes, the built-in MCP server allows AI agents to drive downloads, query archives, and trigger offline speech-to-text.
The system requires PostgreSQL for archiving scraped data and Redis for task management and caching.
See Zekai first in Google