Clean and OCR a scanned PDF
Remove blank pages, improve colors, run OCR, then download the searchable PDF.
What does this template do?
Remove blank pages, improve colors, run OCR, then download the searchable PDF.
This template makes a scanned document clean and searchable: it removes blank pages, strengthens contrast, then runs text recognition (OCR), which adds an invisible text layer under the image. You can then search, copy or index the content.
Pipeline steps
This template automatically chains 5 steps in the Workspace. You can review every setting before running it.
- 1Import a PDF
- 2Remove blank pages
- 3Adjust contrast and colours
- 4PDF OCR
- 5Download the PDF
Why the steps come in this order
Blank pages are removed before OCR so empty pages are not analysed. Contrast is strengthened before recognition, because sharper text is read more accurately.
When should you use this template?
- Make scanned invoices or contracts searchable so you can find them by keyword.
- Copy a passage from a digitised paper document without retyping it.
- Prepare scanned archives before filing them.
Local processing in your browser
When the selected modules support local processing, your files stay on your device while the Pipeline runs.
Settings to check
- OCR language: French by default, English, Spanish, German, Italian, Portuguese or French + English; choose the language of your document.
- Analysis resolution: scale 1.8 by default (about 130 dpi); raise it for small print.
- Contrast: strengthened by a fixed setting (+20), with no option in this template.
Example
- Before
- 3-page scanned invoice with no selectable text.
- After
- 3-page PDF in which you can search for “Total” or the invoice number.
Limits to know
- Recognition can be wrong on a blurred, skewed or handwritten scan: check amounts and names.
- Pick the right language: an English document analysed as French gives more errors.
- OCR is the longest part: expect several seconds per page in the browser.