Editing OCR Text
OCR isn't perfect -- sometimes characters are misread, especially with low-quality photos or unusual fonts. DocSnap lets you correct the extracted text while preserving the original OCR output for reference.
How It Works
- The original OCR text is never modified -- it's always preserved as-is
- Your corrections are stored as a revision alongside the original
- You can always compare the original output with your corrected version
- Each revision is timestamped so you know when changes were made
Making Corrections
- Open a completed document from the Documents tab
- Scroll to the OCR Results section
- Tap the Edit button
- The text becomes editable -- make your corrections directly in the text field
- Tap Save to store your revision
- Tap Cancel to discard changes
Viewing the Original
After saving a revision, an orange Revised badge appears next to "OCR Results" to indicate the text has been modified.
To compare your corrections with the original OCR output:
- Tap Show Original below the OCR section
- The original text appears with an orange border for easy distinction
- Tap Show Current to return to your revised version
This side-by-side comparison helps you verify that your corrections are accurate and complete.
Revision Details
Below the OCR text, you'll see a timestamp showing when the last revision was saved:
Last revised: Feb 21, 2026 at 3:45 PM
Common Corrections
| OCR Misread | Likely Correction |
|---|---|
1 vs l | Number one vs. lowercase L |
0 vs O | Zero vs. letter O |
rn vs m | "rn" misread as "m" or vice versa |
| Missing spaces | Words run together in densely printed text |
| Special characters | Symbols like #, $, @ sometimes misread |
Tips
- Focus on correcting critical data first: reference numbers, amounts, dates, names
- You don't need to correct formatting issues like extra line breaks
- The revised text is what gets used when you export to Word or Excel
- Search indexes are based on the original OCR text, not your revisions