Lab · 06 · Document extractor

Pull structured data out of any document.

Four sample documents — invoice, contract clause, CV, lease — or drop your own image. GPT-4o vision returns a structured JSON of fields with confidence and (where possible) the bounding box of where it found each one. Fields stream into the form on the right as the model identifies them.

Pick or drop a document

PNG, JPEG or WebP up to 8 MB.

Or drop your own document here.

SourceNo document loaded

Pick a sample or drop your own document to start.

Extracted fields0 fields

Fields stream in here as the model identifies them.