FlipCheck: snap a thrift find, get a fair price and a ready listing (Gemma 4 + TabPFN)

Описание и применение

FlipCheck ускоряет процесс оценки и листинга товаров для продавцов на eBay, Vinted и автосалонах, используя модель Gemma 4 для распознавания предмета и генерации описания, а также TabPFN для мгновенной оценки цены на основе живых листингов eBay.

Задача

Resellers spend up to ten minutes per item researching what it is, its market value, and crafting a listing, which reduces profit when handling many items.

Результат

FlipCheck reduces research time from minutes to seconds, enabling sellers to price and list dozens of items per weekend with higher accuracy and consistency.

Рабочий процесс

• Snap photo of item
• Gemma 4 identifies item and outputs JSON (title, brand, model, condition, search query, confidence)
• Query SerpApi for up to 200 current eBay listings matching the search
• Clean listings: remove sponsored, bulk lots, unrelated titles, and outliers
• Feed cleaned listings to TabPFN to predict 25th, 50th, 75th percentile prices for the item’s condition
• Gemma 4 generates eBay title (≤80 chars), description, and item specifics
• Present results to user with copy buttons for each field

Доказательства

FlipCheck does that research from one photo: 1) What is it? An open vision model names the item: brand, model, visible condition, and the search a seller would type.

Источник

It pulls live eBay listings for that search and prices your item’s condition as three numbers: quick sale, fair, and if you can wait.

Источник

An eBay title (80 characters max), an honest description and item specifics, each with a copy button.

Источник

Gemma 4 gets the photo and must answer in a strict JSON schema: title, brand, model, condition, search query and confidence.

Источник

TabPFN is a tabular foundation model. It fits a small table in one forward pass, with no training run.

Источник