Snap a photo, get a listing: a local Gemma 3 listing generator for secondhand clothes

Description and use

The system lets Mano en Mano sellers quickly create clothing listings by uploading 1‑4 photos and receiving an auto‑filled 12‑field form generated locally by Gemma 3 via Ollama.

Problem

Sellers on Mano en Mano must manually fill 12 fields for each garment, making bulk listings slow and discouraging listings.

Outcome

The system reduces listing time from minutes per garment to seconds, enabling sellers to list dozens of items efficiently while keeping data privacy and avoiding API costs.

Workflow

• Upload 1‑4 photos of a garment
• Gemma 3 (gemma3:4b) runs locally in Ollama and analyzes the images
• The backend POST /generate-listing returns a JSON listing matching the marketplace’s schema
• The frontend displays an editable form pre‑filled with the generated data
• Seller reviews, corrects any fields, and publishes the listing

Evidence

Upload 1 to 4 photos of a secondhand garment and get an editable product listing with 12 fields, generated locally by an open-weight vision model.

Source

Gemma 3 (4B, vision‑capable), open weights.

Source

On my laptop, each garment takes roughly 15 to 30 seconds (the first request is slower because it loads the model).

Source

Sellers’ photos go from the browser to my own server to a model running on the same machine. They never reach a third party.

Source