Chat Screenshot Cleaner

Yes. Adding, dropping or pasting screenshots starts text recognition for the new images. You can watch the progress or select Stop. Existing screenshots keep their detected text and edits when you add more images. The tool does not apply covers automatically; you choose what to hide. Use Scan text again to rescan the whole batch.

Dark backgrounds, coloured message bubbles, small text and compression can hide lettering from the reader. The cleaner enlarges small screenshots, adjusts light or dark contrast and checks coloured bubbles in a second pass. If it still finds no words, change Screenshot appearance and scan again using the original full-size screenshot. The reader currently uses English; other languages may need manual covers.

No. Screenshots, detected text, avatar comparisons and exports are processed in your browser tab. The text reader and face detector download their files from ToolsNow when needed. Your images and recognised text are not sent to a processing service.

Yes. Scan the screenshots, enter the name exactly as the reader found it, and apply a cover to its matches across the batch. Matching ignores letter case and covers whole recognised words. You can replace the covered area with a consistent label such as Person A. Misspelled or unread text needs manual review.

No. You choose names and usernames to hide. The contact suggestions look for email and phone-number patterns, which can also match ordinary numbers. Face detection finds possible faces without identifying people. Always review the result yourself.

Draw a cover tightly around an avatar and select Find this avatar in all screenshots. It searches for similar copies of that image at a similar scale. This is image matching, not identification of a person. Changed avatars, crops, sizes, compression and themes may prevent matches or produce false matches.

The export replaces covered pixels with an opaque colour and, optionally, your anonymous label. It does not attach original images, editing layers or recognised text. The output is a new flattened image with no metadata copied from the originals. Visible details elsewhere in the screenshot may still identify someone.

You can open screenshots from WhatsApp, Messenger, Discord, iMessage, Slack or other apps; no account connection is needed. The first text reader uses an English model. Results vary with language, font, image quality and theme. Manual covers and stitching work independently of text recognition.

Screenshots join vertically in the order shown. The tool can suggest strong overlapping sections, and you can adjust top and bottom trims for each image. Remove repeated app headers first and review every join. Images shrink to a common width without upscaling. Separate PNGs in a ZIP preserve each image’s width.

A batch supports up to eight images, 15 MB per file, 12 million pixels per image and 32 million input pixels in total. Input images and joined exports allow up to 16,000 pixels per side. A joined export is limited to 24 million pixels. These limits help keep phone browsers usable. Choose a smaller output width or separate images if the joined image is too large.