SellSnap AI

SellSnap AI turns one product photo into a marketplace listing, flags details AI cannot verify, and lets sellers confirm missing information.

Comment

Inspiration

What it does

How we built it

Challenges we ran into

Inspiration

Creating a marketplace listing manually can take time. Sellers need to write a title, choose a category, describe the product, add specifications, and make sure important details are not missing.

I wanted to create a tool that could make this process faster while also avoiding one common AI problem: inventing product information that cannot actually be verified from a photo.

What it does

SellSnap AI turns a single product photo into a marketplace listing draft.

The AI analyzes the image and generates:

  • Product title
  • Marketplace category
  • Short description
  • Visible product attributes
  • Listing Quality score
  • Product Recognition confidence
  • Missing details that cannot be verified from the image

Instead of guessing unknown information, SellSnap AI asks the seller to confirm details such as price, condition, brand, model, material, size, or capacity.

After the seller adds the missing information, SellSnap AI generates a final ready-to-copy marketplace listing.

How I built it

I built SellSnap AI as a responsive web application using React, TypeScript, JavaScript, and Floot.

The application uses AI vision to analyze product photos and generate structured listing information.

I also added validation logic to prevent incomplete or ambiguous seller data. For example, a price must include a currency such as "12 TJS", and measurements such as capacity must include a unit such as "350 ml".

The final listing combines information detected from the image with facts confirmed by the seller.

Challenges I faced

One of the biggest challenges was preventing the AI from confidently inventing information.

A photo may show that an object is a white mug, but it may not reveal its exact capacity, material, condition, brand, or model.

To solve this, I separated visible AI-detected information from seller-confirmed information.

I also worked on making the Listing Quality score more meaningful by reducing the score when important information is missing or incorrectly formatted.

Accomplishments that I'm proud of

I created a complete photo-to-listing workflow:

  1. Upload or capture a product photo
  2. Generate an AI draft
  3. Review visible product information
  4. See what the AI cannot verify
  5. Add seller-confirmed details
  6. Generate the final listing
  7. Copy the listing for use on a marketplace

I am especially proud that SellSnap AI does not simply generate text. It also checks whether important information is missing before the listing is ready.

What I learned

While building SellSnap AI, I learned more about working with AI vision, structured AI outputs, input validation, responsive UI design, and designing AI systems that clearly separate predictions from confirmed facts.

I also learned that a useful AI product should not only generate information, but should communicate uncertainty when it does not know something.

What's next for SellSnap AI

In the future, I would like to add support for more product categories, multiple product photos, automatic language translation, marketplace-specific listing formats, and direct integrations with e-commerce platforms.

My goal is to make SellSnap AI a practical assistant that helps sellers create accurate product listings faster.

Built With

Share this project:

Updates

Submission history