OCR turns a scanned image of a page into text you can search and copy. Here is what it does, where it falls short, and how it differs from indexing and AI data extraction.
OCR stands for optical character recognition. A scanner takes a picture of a page, and a picture has no words in it, only pixels. OCR software looks at the picture, recognises the letters and adds a hidden layer of text behind the image. The page looks the same, but you can now search it, select text and copy it.
Without OCR, a scanned archive is a set of pictures that you can only find by file name. With it, you can search every page for a name, an invoice number or a phrase, and get the right document back. That is most of the value of scanning in the first place.
Find any word across thousands of pages.
Select text from a scan without retyping it.
Text can be read by screen readers.
OCR depends on the quality of the original. Faded thermal paper, crumpled pages, stamps over text, small print and handwriting all reduce accuracy. Good scanning, such as straightening pages and cleaning up images, helps, but it cannot recover what the paper no longer shows.
Accuracy also matters more for some uses than others. Searching an archive tolerates the odd missed word. Extracting an invoice total or a date into a system does not, so those results need checks.
These three are often confused. OCR makes pages searchable. Indexing names and organises the files by what they contain. Data extraction pulls specific details, such as dates and amounts, into fields in your system.
OCR is included in our standard scanning. AI indexing is optional, and document conversion and data entry covers extraction into your system. See our prices.
With us, yes. OCR is included in standard scanning, so every page is searchable. Some providers charge per page for it, so check what is included before comparing quotes.
No. OCR turns the picture of a page into text. AI indexing reads that text to work out what the document is, then names it and tags it with the details that matter, such as a date or reference. OCR comes first, and AI builds on it.
It depends on the original. Clean, printed pages read well. Faded paper, skewed scans, stamps over text and handwriting all reduce accuracy. A sample of your own documents is the only honest test.
Standard OCR is built for printed text and handwriting is much harder. If handwriting matters to your project, ask us to test it on a sample before quoting.
Tell us roughly how many boxes and what is in them, and we will come back with a quote. Collection is quoted separately.