Secondhand searches are often less about finding any listing than repeatedly checking whether a specific, imperfect match has appeared. This project automates that routine and adds a useful distinction: users can describe what they want in natural language, while image-aware AI helps judge listings beyond exact keyword matches.
What Sets It Apart
- AI screening alongside simple rules: Use natural-language criteria for model-based analysis, or keep a task to keyword and price rules. That makes it possible to apply nuanced preferences without requiring AI for every search.
- A monitoring workflow, not just a scraper: Multiple tasks can run with their own filters and schedules, and results can trigger notifications. The practical benefit is less manual refreshing and a clearer path from a match to an alert.
- Designed for ongoing operation: A web interface, SQLite-backed records, account management, and proxy rotation bring the pieces of a self-hosted service together. This is useful for long-running searches, but also means deployment and account upkeep are part of the deal.
Who It’s For—and the Trade-Off
Great fit if you regularly hunt for particular secondhand items and can supply an OpenAI-compatible API that accepts images, manage a server, and use a Xianyu login. Look elsewhere if you want a hosted service, do not want to provide account login state, or need guaranteed platform access: browser automation can encounter login expiry and anti-abuse controls, and the project asks users to respect Xianyu’s terms and request limits.
