OCR scanned files
Recognize English text in an image or image-only PDF.
Open OCR scanned filesWhen to use it
A scanned letter looks readable but copy and paste produces nothing. Run OCR and export TXT for text, DOCX for editing, or a searchable PDF for finding words.
Recover text from a clear English scan
Use the practice scan containing “Delivery note”, reference FR-2048, and quantity 12.

Filerivo-created demonstration files. No real personal or customer data. Settings below are a recipe, not an automatic preset.
- Choose the practice scan PNG and open Recognize and review text.
- Compare recognized text with the image. Correct the reference and quantity if recognition differs.
- Choose reviewed TXT, enable reviewed export, and choose Apply, Preview, then Download.
- Expected result
- Editable English text containing the source wording after your corrections. Exact spacing from the scan is not preserved.
- Check before sharing
- Check FR-2048 and 12 character by character. OCR accuracy is not guaranteed even on a clear sample.
How to do it
- Choose a PDF, JPG, PNG, or WebP file in OCR scanned files.
- Choose TXT, Word, or searchable PDF output.
- To correct recognition before downloading, open Advanced OCR review editor and select Recognize and review text. Compare each source page with its editable text. Join single line breaks, restore recognized text, or undo the last helper if needed.
- Choose reviewed TXT or Word and enable Download reviewed text instead of quick OCR output. Choose Apply, Preview, then Download before changing files or tools; those changes clear your edits. Reviewed edits do not update searchable PDF output.
- Run OCR, then open the result and proofread names, dates, amounts, and small print against the original.
Apply, preview and download
Choose Apply after selecting files and setting your options. Complete any required advanced recognition, review or workflow steps first. Apply prepares a result without downloading it. Choose Preview to review that prepared output. Review generated TXT or Word text, or navigate the pages of a searchable PDF. Word is shown as text rather than a faithful Word layout. Choose Download to save the same generated file.
Preview and Download become available after Apply. Changing files, settings or visual edits requires Apply again. Content preview supports up to 100 MiB per output or archive entry; text and Word content have a 2 MiB preview limit and long text display is truncated. Keep originals and inspect the full download in its intended application.
Use the advanced controls
Advanced controls are shown dimmed before a file is selected. Add a supported file to activate them. Where a Load preview or Extract button is shown, use it first; enable the tool's advanced-download checkbox to apply those settings. Changing files or tools clears unsaved edits.
Recognize and review text loads source pages beside editable recognized text. Correct names, numbers, and punctuation on each page. Join single line breaks changes wrapping; Restore recognized text discards corrections on that page, and Undo last helper reverses a helper action.
Reviewed output can be TXT or Word when Download reviewed text instead of quick OCR output is enabled. Searchable PDF remains a separate quick output and does not include review corrections. Cancel review recognition stops after the current operation; download edits before changing files or tools.
Check the preview and download
Previews render at higher pixel density and are scaled to fit the page. They do not restore detail missing from the original, and their display size does not change export dimensions. For encoded-output comparisons, check both visible detail and the reported size. Always open the downloaded file for a final check.
Know the limits
OCR currently recognizes English only. PDFs are limited to 30 pages. Reviewed exports support up to 100,000 characters per page and 1,000,000 total. Handwriting, low contrast, skew, and unusual layouts can reduce accuracy. Word contains text rather than the original design. Searchable PDF remains a separate quick output: the PDF is rebuilt, its appearance can change, and review edits are not applied to its text layer. Cancel stops review recognition after the current operation finishes.
Troubleshooting
Recognition takes a long time
The first run loads recognition resources. Keep the tab open and try a smaller, clear image if device memory is limited.
Letters or numbers are wrong
Use a straight, well-lit, higher-resolution scan. Proofread and correct reviewed TXT or Word before downloading.
Reviewed edits are absent from searchable PDF
Reviewed corrections apply only to TXT or Word exports. Searchable PDF is a separate output path.
Still stuck? Keep the original and note the tool, file type, approximate size and error message. Contact Filerivo with those details; do not email confidential files.