We Searched eBay for 237 Recalled Products. Only 45 of 12,562 Listings Named the Model.

Developers @dronzer2code and @soumyadeepdey have introduced RecallDesk, an open-source tool designed to help resale shops identify recalled products. By analyzing U.S. Consumer Product Safety Commission (CPSC) data, the team discovered that identifying information is frequently buried in product descriptions rather than titles, making it difficult for automated systems to flag unsafe items on platforms like eBay. RecallDesk uses a fine-tuned Qwen3.5-4B model to parse recall notices and compare them against a shop's inventory, generating a weekly 'pull list' for staff. The project emphasizes a 'rules-first' approach, combining deterministic logic with LLM-driven classification to handle ambiguous cases. The tool, built with Next.js and Temporal, aims to bridge the gap between federal safety notices and the realities of second-hand retail, providing a cost-effective way for small businesses to ensure compliance and consumer safety.