Retro games · Vinyl · Books · Trading cards

Photograph your collection into a database.

Thousands of items in boxes, none of them searchable? Photograph them one by one with QR-code dividers between boxes, and Visual Cataloguer turns the camera roll into a browsable inventory — AI identifies each item, tracks where it's stored, and checks what it recently sold for on eBay.

Download the desktop app Other ways to run it

 

All releases →

Local-first. One SQLite file, images included — no cloud, no accounts.

How it works

  1. 1

    Print QR dividers

    One per box or shelf — BOX-1, SHELF-A3. Hand-written labels work too.

  2. 2

    Photograph everything

    Divider, then one photo per item. Cover the lens to end a box. Multiple cameras merge by timestamp.

  3. 3

    Process the folder

    AI reads the dividers and identifies every item — title, platform, region, completeness, condition.

  4. 4

    Search, price, list

    Browse the web UI, check recent eBay sold prices, and mark items as listed when they sell.

Built for big boxes of stuff

🤖 AI identification

Claude in the cloud or LLaVA on your own machine via Ollama — titles, platforms, regions and completeness from a single photo.

💰 eBay price research

One click pulls recent sold listings for an item and shows the low, median and high — with the sales that back it up.

📦 Location tracking

Every item remembers which box or shelf it came from, so you can actually find it again when it sells.

📷 Camera-friendly

JPEG, PNG, TIFF and ~20 RAW formats. Auto-crop, deskew and rotate. Duplicate photos are detected and skipped.

🔍 Searchable web UI

Filter by platform, box, completeness or review status. Edit titles, crop images and generate eBay descriptions inline.

📤 Yours to export

CSV, JSON, or images organised by location — plus a REST API. Everything lives in one portable SQLite file.

Three ways to run it

💻 Desktop app

Download, unzip, run. No Python, no Docker — your database lives in your user folder.

Download

Unsigned builds: on macOS right-click → Open the first time; on Windows choose "More info → Run anyway".

🐳 Docker

For a home server or NAS — the whole household can browse the collection.

curl -O https://raw.githubusercontent.com/retroverse-studios/visual-cataloguer/main/docker-compose.yml
docker compose up -d

Pulls ghcr.io/retroverse-studios/visual-cataloguer. Web UI on port 8000, database in ./data/.

🐍 Python package

The full CLI for batch processing, plus the same web UI via viscatalog serve.

pip install visual-cataloguer[web]
viscatalog process ./photos
viscatalog serve

Python 3.11+. Offline OCR mode needs Tesseract.