Extract data from documents with AI (OCR)
The AI reads the image of an invoice, receipt or bill and extracts the structured fields (vendor, date, total, currency, document number and line items). Attach the image to a record and click Extract data (AI); the result shows up in a readable panel with a button to copy the JSON.
Document extraction uses the same system AI to read the image of a business document (invoice, receipt, bill) and return the fields in a structured format: vendor, date, total, currency, document number and the list of items.
How to use it
- Open the record and go to the Attachments section.
- Upload or select an image of the document (PNG, JPG, WEBP or GIF).
- Click Extract data (AI) next to the attachment.
- Review the fields in the Extracted data panel and use Copy JSON if you want to take the result elsewhere.
💡 Each extraction shows a confidence level (High, Medium or Low) and a short note. Always double-check the values before using them, especially on low-quality photos.
In this first version, extraction works with images. PDF documents are coming soon. Extraction uses the system AI, so the AI must be configured; usage is metered like all other AI usage.