OCR (optical character recognition)
OCR (optical character recognition) is the technology that converts the text visible in an image — a scanned page, a photographed receipt, an image-only PDF — into machine-readable text that software can search, copy, and analyze. Without OCR, a scan is just pixels: a computer can display it but has no idea what it says.
Why scans are the problem files
Documents arrive in two forms. Native digital files (a Google Doc, a PDF exported from an invoicing tool) carry their text internally — search finds them, software reads them directly. Scans and photos don't. A passport photographed with a phone, a paper invoice run through a scanner, a receipt snapped at a restaurant: all text-as-image. These are precisely the files that pile up with names like scan_0042.pdf or IMG_4523.HEIC, because nothing about them tells you — or your tools — what's inside.
OCR closes that gap. Run on the receipt photo, it produces the merchant name, the date, the line items, the total — as text. Modern OCR handles skewed photos, multiple languages, and mediocre scan quality far better than the technology's reputation suggests, though it remains probabilistic: a crumpled thermal receipt or a low-light photo can still come back partly garbled.
OCR in a sorting pipeline
For file organization, OCR is the first of two steps:
- OCR extracts the raw text from the scan or photo.
- Classification interprets it: this is an invoice, from supplier Acme, dated 2026-06-12, for 840 €.
The second step is what an AI file organizer adds on top. Together they turn an unreadable filename into a placed, renamed file:
scan_0042.pdf → Invoices / 2026 / 06 - June / acme-2026-06-12.pdf
This pipeline is built into Sorters for Google Drive files: when you run a sort, each picked file — PDF, JPG, PNG, HEIC and more — goes through OCR and AI classification so the rename-and-file rules can use what the document actually says, not what the scanner named it.
Do you need OCR?
Only if image-form documents are part of your mess. A Drive full of native Google Docs and exported PDFs is searchable as-is — Drive even applies some OCR to scans in its own search index, which is why searching a supplier name sometimes surfaces an unnamed scan. What Drive's built-in OCR won't do is act on the result: it finds the scan, but renaming and filing it remains manual or remains a tool's job.
Related
How to organize invoices in Google Drive shows the most common use of this pipeline end to end, and automatic file sorter covers where OCR-based sorting sits among the other sorting approaches.