ai-goofish-monitor

This project is an automated monitoring system for Xianyu (informally known as Idle Fish), China's largest second-hand marketplace run by Alibaba. Instead of manually refreshing the site to check for new listings, you set up tasks that watch for specific items around the clock and alert you the moment something matching your criteria appears. It is written in Python and uses a browser automation library called Playwright to navigate Xianyu the way a human would, without a formal API.

You control everything through a web interface that runs on your own computer or server. From there you create monitoring tasks by describing what you want in plain language and letting the built-in AI figure out the details, or by typing keywords directly. Each task can target specific price ranges, filter by whether the seller ships or only meets locally, restrict results to a particular city or district, and apply its own AI prompt to judge whether a listing is actually worth buying. Multiple tasks can run at the same time, each with its own settings and its own bound account.

When a listing passes all your filters and the AI rates it as worth your attention, the tool sends you a notification. You can receive alerts through several services including Telegram, Bark, WeChat Work, ntfy.sh, or a custom webhook. The notification carries the item details the tool gathered from the listing page.

For reliability, the system lets you manage multiple Xianyu login sessions and rotate between them so that no single account gets flagged for excessive activity. It also supports a proxy pool with automatic rotation and retry on failure. All results, price history, and logs are stored in a local SQLite database.

Running it is straightforward via Docker: a single command pulls the image and starts the server with a built-in Chromium browser, so you do not need to install a separate browser on the host. A web interface is then available at port 8000. The project is released under the MIT license and is described by its authors as intended for personal learning and research rather than commercial scraping at scale.