Extract & annotate
art historical metadata
Upload a ZIP of scanned catalogue pages. The pipeline uses Google Vision for cropping and Gemini for structured metadata extraction, then presents results for human review.
📁
Click to browse or drag & drop
.zip archive or a single .jpg / .png / .tif image
Initialising…
0 / 0
Processing — this may take a few minutes
Each page is being analysed by Google Vision and Gemini in turn. The log below updates in real time.
01
Vision Crop
Google Vision detects and crops artwork regions from raw scan pages.
02
Gemini Extract
Gemini reads each page and extracts all metadata fields as structured JSON.
03
Human Review
Browse results image-by-image. Edit any field inline and export a clean CSV.