Realist: an open-weight listing auditor that finds the mismatch before your buyer does
说明与用途
Realist 自动审核 eBay 和 Etsy 上的商品列表,检测标题、描述、规格与照片之间的不一致,并在买家发现错误之前提供可操作的纠正建议。
问题
Sellers on eBay or Etsy often publish listings with small but costly mismatches between titles, descriptions, specs, and photos, leading to returns, refunds, or negative reviews.
结果
Realist detects text conflicts with 100% confidence, image mismatches with up to 80% confidence, and flags possible omissions or unverified claims, enabling sellers to correct listings before buyer complaints.
工作流
- Paste title, description, and spec list into Realist
- Upload up to eight photos
- Realist extracts and normalizes fields via plain Python
- Vision model analyzes photos for observable attributes
- Python compares text and image observations
- Realist outputs findings with confidence and next steps
证据
Realist returns findings. Every finding has the exact fields or photo involved, a confidence, and a next step.
Plain Python does the comparing. It extracts Bluetooth version, IP rating, weight, quantity, battery capacity and color from the title, description and specs, normalizes them, and checks them against each other.
A vision model only reports what it sees. It returns structured observations per photo (colors, visible accessories, legible text, item count) and never judgments.
Seller data stays on the seller's machine. Realist runs Qwen2.5-VL-7B through a local Ollama server (default qwen2.5vl:7b).
- 记录日期: 2026-10-14T05:41:05.226Z