StockEntry Guard: Local Gemma-Powered Product Onboarding for Small Retailers

说明与用途

StockEntry Guard 使用本地 Gemma 3 模型从商品图片中提取数据，检查重复项，并让零售商手动批准将商品添加或更新到本地 SQLite 库中。

问题

Small retailers struggle with duplicate product entries when catalogs grow, leading to messy inventories.

结果

StockEntry Guard reduces repetitive data entry and prevents duplicate products by combining local AI extraction with deterministic duplicate detection, all under human approval.

工作流

• Import existing inventory CSV
• Upload product photo
• Extract product details locally with Gemma 3 via Ollama
• Review and edit extracted fields
• Compare candidate against existing catalog for duplicates
• Choose to create new product, update existing, or discard
• Save approved product to local SQLite database
• Export updated inventory to CSV

证据

StockEntry Guard helps a retailer: Import an existing inventory CSV. Upload a product photo. Extract product details locally with Gemma 3. Review and edit the extracted fields. Compare the product against the existing catalog. See whether there is no match, a possible match, or a likely duplicate. Create a new product or update an existing one only after explicit human approval. Export the updated inventory back to CSV.

来源

The demo shows the complete retailer workflow: Import an existing inventory CSV. Upload a product image. Extract structured product details locally with Gemma. Review and edit the extracted fields. Check the product against the existing inventory for possible duplicates. Approve the product. Save it into the local inventory database. Export the updated inventory back to CSV.

来源

StockEntry Guard runs Gemma 3 locally through Ollama for product-image extraction.

来源

The extracted result is never written directly to inventory. The retailer reviews the fields first, then runs StockEntry Guard's deterministic duplicate check.

来源