>-

The Challenge of Extensive Listing Work

In the reuse and retail businesses, the most time-consuming and skill-dependent task is "listing work (research, title creation, and product description writing)."

Manually researching "What category does this product belong to?", "What is the market price?", and "What keywords are likely to be searched?" each time was the biggest bottleneck hindering business scaling.

Solution: Automatic Listing Data Generation Bot using Telegram × Gemini

To fundamentally solve this problem, I developed a unique AI bot (Yahoo! Auctions AI) from scratch that "returns all the necessary information for listing simply by sending images from a smartphone."

【Main Features of the System】

1. 【Image Grouping Engine】: When images are sent continuously from a smartphone via Telegram, it waits for a few seconds and automatically recognizes them as "one product (group)."

2. 【JSON Curation AI】: Images and a simple comment (e.g., "Junk, operation unconfirmed") are passed to the AI, which accurately generates the following data in JSON format.

• Optimized [Title]
• Yahoo! Auctions [Category]
• Attractive [Product Description]
• Market-based [Price]
• [Research Points] for Searchability

3. 【Database State Management】: The generated data is sent to the smartphone chat screen in an easy-to-read format, and the status (processing, completed, etc.) is strictly managed in the underlying SQLite database.

Result: Listing time reduced to 1/10, eliminating reliance on individual expertise

This system reduces the creation of listing data, which previously took several minutes to over ten minutes per item, to just "seconds."

Even part-time staff without special knowledge can now create listing data with professional-level SEO (search engine optimization) simply by taking photos and sending them to the bot, dramatically improving the business's profit margin and turnover rate.

I don't just develop tools; I propose essential business improvements that fundamentally eliminate the "hassle" of on-site work through system implementation.