How to extract invoice data with OCR
Invoice OCR (optical character recognition) reads text from scanned invoices and images. TouchInvoice uses the extracted text to fill in fields such as the supplier, invoice number, dates, totals and line items. You can review and edit those fields before exporting the data.
1. Upload the invoice
Sign in with a verified email and upload a PDF, JPG, PNG, WebP, BMP or TIFF file. Each file can be up to 10 pages and 10 MiB. For a scanned invoice, use a clear image that includes the whole page, especially the totals near the bottom.
2. Check the extracted fields
Open the processed invoice in the review workspace and compare the data with the original. Start with the invoice number, date, currency, tax and amount due. If the invoice has line items, check that each description, quantity and amount appears on the correct row. Use the original invoice to verify any payment instructions.
For example, if the original shows NW-1048 and £240.00, confirm that the extracted invoice number and total match before using the result. Correct any missing or misread characters in the workspace.
3. Export the reviewed data
Export CSV to open the invoice data in a spreadsheet, JSON to use it as structured data in another tool, or text to keep a readable copy of the extraction. Keep the original invoice alongside the export so you can verify the values later.
Compare document allowances · Learn how the TouchInvoice workspace handles your invoices