OCR Not Working? Common Problems and Fixes

OCR rarely fails silently for no reason. When the output is garbage, blank, or full of wrong characters, it is almost always the input image or a setting, and both are fixable.

The output is blank or empty

If you get nothing back, the tool likely could not find any text at all:

The image has no actual text, or the text is too small to detect. Crop tighter to the text and re-upload.
Contrast is too low, so letters blend into the background. Increase brightness or contrast before uploading.
The file is corrupt or an unsupported format. Re-save it as a PNG or JPG and try again. Our guide on which image formats work best for OCR covers the safe choices.

The text is garbled or full of wrong characters

This is the most common complaint, and the usual causes are:

Blurry or out-of-focus images. OCR needs sharp edges. See whether OCR can read blurry images for what is recoverable.
Low resolution. Tiny text simply lacks the detail to recognize. Aim higher, as our best resolution for OCR guide explains.
Skewed or rotated pages, which break the line detection. Straighten and rotate upright first.
Wrong language selected. If the engine expects English but your text is German, accented letters come out wrong. Pick the matching language.

Handwriting comes out as nonsense

This one is expected, not a bug. Tesseract-based OCR is built for printed text and struggles badly with handwriting, especially cursive. Block capitals fare a little better. For realistic expectations, read how accurate handwriting OCR is, and try our handwriting to text tool knowing the results will need correcting.

Tables and columns get scrambled

If a table comes back as one jumbled run of text, the engine read across columns instead of down them. That is a known weak spot. Our guide on OCR and tables explains the workarounds, such as cropping one column at a time.

A quick checklist before you give up

Is the image sharp and in focus?
Is it at least 300 DPI or a high-resolution photo?
Is the page straight and upright?
Is the contrast strong, dark text on a light background?
Did you select the correct language?
Is the format a standard PNG, JPG, or PDF?

Fixing the input fixes most failures. For the full set of techniques, see improve OCR accuracy.

Common questions

Why does my screenshot OCR perfectly but my photo does not?

Screenshots are pixel-perfect and high-contrast; photos add blur, glare, angle, and uneven light. Flatten and brighten the photo, or re-take it straight-on.

The tool says "no text found" but I can clearly see text.

The text is likely too low-contrast or too small for the engine to detect. Crop closer and boost contrast, then re-upload.

Is it my file or the tool?

Test with a clean printed screenshot. If that works, the problem is your original image, not the tool.

Try a clean source

Most "OCR not working" cases vanish with a sharper, straighter, higher-contrast image. Fix the input, then run it back through image to text, free, no account, files deleted after processing, and the results usually fall into place.