How to unblur text in an image
Blur removes information from a picture, so no tool truly puts it back. What you can do is find a sharper copy, improve what is there, and let OCR read the recovered characters.
First, understand what blur actually did
Blur is not a filter sitting on top of the words. It spreads each point of the image across its neighbours, and the detail that separated an “e” from a “c” is averaged away. That is why the honest question is not “how do I unblur text” but “how much readable signal is still in this file, and how do I get at it”.
This matters for expectations. Slight motion blur or a mild out-of-focus shot often cleans up enough to read. Heavy blur, a tiny thumbnail scaled up, or text deliberately blurred to redact it is gone, and any tool promising to reconstruct it is guessing at characters rather than recovering them.
Look for a sharper copy before you edit anything
The fastest fix is usually not an editor at all. Messaging apps, social platforms, and email clients all recompress images, so the copy in front of you is frequently not the sharpest one that exists.
- Check the original camera roll or Photos library instead of the shared version.
- Ask the sender to resend the file as a document or attachment rather than an inline image.
- If it came from a website or PDF, download the source file instead of screenshotting it.
- On a phone, take a fresh photo of the paper rather than photographing a screen showing it.
Improve the image you do have
When there is only one copy, small adjustments in any photo editor can raise the contrast between the letters and the background. Increase contrast, reduce brightness slightly so pale strokes darken, and apply a modest sharpen. Stop as soon as the edges start to show halos, because over-sharpening creates artefacts that OCR reads as extra characters.
Crop tightly around the passage you need and enlarge that crop rather than the whole page. A single paragraph blown up to fill the frame gives a recognition model far more pixels per character than a full page where the text is only a few pixels tall.
Converting to greyscale helps when coloured text or a busy background is competing with the words. It removes colour noise without touching the shapes you care about.
Use OCR to read the recovered text
Once the crop is as clean as you can get it, OCR is the step that turns the pixels back into characters you can copy. Upload it to unblur text, which is tuned for exactly this case, or use photo to text if the source is a camera shot of paper.
OCRphoto returns editable plain text rather than a repaired picture. If the words come back correctly, you have what you needed; if there is not enough detail left, it says so instead of inventing a transcription that reads convincingly but is wrong.
Unblurring text on an iPhone
On iOS you can do the whole improvement pass without extra apps. Open the image in Photos, tap Edit, then use Brilliance, Contrast, and Sharpness. Pinch to crop around the text before you export, and share the edited copy at full resolution rather than screenshotting the editing screen.
iOS Live Text can also lift words directly out of a photo when the text is sharp enough. If it cannot find anything, run the edited crop through OCR instead, which often succeeds on marginal images where the built-in selection gives up.
When the text is not coming back
Text that was blurred on purpose to hide a password, an address, or an account number is designed not to be recoverable, and treating a guessed reconstruction as fact is how people end up acting on wrong information. The same applies to a heavily pixelated screenshot: a model can produce confident-looking output that is simply invented.
If accuracy matters, verify every recovered name, number, and date against another source before you use it. When there is no second source, ask for the information again rather than trusting a reconstruction.