Edit text in images
and screenshots.

Open a PNG or JPG and select a detected text region. Chinese and English recognition runs automatically in your browser. Type directly on the image or in the Selected text box; both update the same preview. Review the result, then save the edited image as a PNG. Use clear text on a plain background.

Click the detected text region and replace only the characters you need to change. You can also drag to select a smaller area or a missed word. The screenshot text editor keeps the matched font size, weight and natural character widths rather than automatically stretching or shrinking the letters. Longer text shows an overflow warning when it needs more space and cannot cross protected content or a detected table-cell boundary.

This online beta accepts clear PNG and JPG images up to 10 MB and 20 million pixels (20MP). PDF editing is not yet enabled.

Yes. Browser recognition uses a Chinese/English model for Chinese, English-only and mixed-language screenshots. Recognition and font matching are automatic, with optional adjustments. Chinese font alternatives include 中文黑体 and 中文宋体 in regular and bold. Compare the recognized words, punctuation and numbers against the source image before saving; OCR can make mistakes.

Yes, for clear screenshots with straight table borders. Detected grid lines are protected while you edit words or numbers within one cell. A replacement cannot cross a detected cell boundary. The result is an image exported as PNG; your original spreadsheet and formulas are unchanged. Broken, slanted or text-overlapping borders require review.

Letters keep their natural widths and kerning. The same letter count does not mean the same width: iii is much narrower than WWW. Replacements keep the selected size and weight; use shorter text or manually expand the selection into empty space if a replacement overflows.

Matching compares the source pixels with available system fonts and bundled alternatives. It is approximate; review the font, digits and punctuation before saving.

Yes. Drag an empty area or choose Add text. Nearby source text supplies an estimated style. Existing words, graphics and detected table borders remain protected.

PDF editing is not yet enabled in this online beta. The local PDF editor is retained for a later backend deployment.

A longer word, different digits or a wider font may exceed the available space. The editor does not automatically shrink or stretch the text. Use shorter wording or expand the area into empty space; optional size and spacing controls are in Advanced adjustments. Protected neighboring content and table-cell boundaries still limit the edit.

An edited PNG at the original image width and height, generated in your browser. Your original image, recognized text and finished PNG are not uploaded.

The current image and draft stay in browser memory. Clear the image or close/reload the page to discard the session. Only public recognition models may be cached. No persistent image draft or server file history is created.

Yes. PNG or JPG files can be up to 10 MB in file size and 20 million pixels (20MP) in resolution. For example, 5,000 × 4,000 pixels is 20MP. These are two different limits. Zoom in to work on small text, then download the edited PNG at its original pixel width and height. Review the preview before saving.