GPTypo now runs in your browser

You don't write in a text box on our website. You write in Gmail, in your CMS, in a Notion doc, in the form where you're about to hit Submit. Until now, checking that writing meant copying it out, pasting it into the scanner, reading the result, and pasting your edits back. That round trip is friction, and friction means you skip the check on exactly the drafts that needed it.

So we put the scanner where the writing happens. The GPTypo Chrome extension is live, free to install, and it scans any text you select on any page.

How it works

Select text on any page. A small "Scan with GPTypo" button appears next to your selection. Click it (or use the right-click menu, or open the side panel directly) and you get the same thing you'd get on the site: a score from 0 to 100, with every sentence scored on its own and plain-language reasons for each flag.

Click a flagged sentence and you get a rewrite tuned to your tone. On writeable fields, one click replaces the sentence in place. On read-only pages, you copy it out. No tab switching, no losing your place in the draft.

Where it works

The extension runs everywhere you'd want it to. We built dedicated handling for the surfaces writers use most:

• Gmail, LinkedIn, Substack, Notion. Writeable fields, so you can scan and replace inline.
• Reddit, and any read-only page. Scan the text, get your rewrite, copy it where you need it.
• Anywhere else. If you can select it, you can scan it: web forms, your school's LMS, a Google Doc, a comment box.

It does not read pages on its own. It only ever acts on the text you deliberately select and scan. That text goes to our API for scoring and is never sold. The full detail is on the privacy page.

Free, with the same Pro you already know

The extension is free. The free tier runs our fast heuristic scanner on up to 500 words at a time, which is plenty for an email, a post, or a paragraph you're unsure about.

If you have a Pro account, sign in once and the extension unlocks verified scoring from our independently benchmarked detection model, unlimited rewrites, and shareable proof reports. One account, the web app and the extension, the same history. Nothing new to manage.

The best time to catch a false positive is before you hit Submit, not after someone emails you about it.

Why this matters for false positives

Roughly 1 in 20 pieces of genuine human writing gets flagged as AI. The patterns that trip detectors (uniform sentence length, a stack of furthermores, flat formality) are easy to introduce without noticing, especially in the polished, careful writing you do for work or school. Those are exactly the signals detectors weight, and they have nothing to do with whether you actually used AI.

Scanning at the point of writing means you see those signals while you can still do something about them. You're not defending finished work after a flag. You're adjusting a sentence or two before anyone else reads it, in the same voice, making the same argument, just clearer in the spots that were tripping the detector. If you do get flagged anyway, we wrote a separate playbook for handling it.

Get it

Add the extension from the extension page, select a sentence in whatever you're writing, and click Scan. The signal tags will show you exactly why it got flagged, and the suggested rewrite gives you a starting point you can take or edit. Feedback and feature requests go to @gptypo.