Paste text. Click an image. Get a second opinion.

What it checks

• Text. Paste a headline, tweet, paragraph or a short article. The AI judges internal consistency, sensational phrasing, fabricated-sounding numbers, and surfaces up to five concrete claims a cautious reader should re-verify with a primary source.
• Images on any page. Click “Check image on page”, then click any image. The model looks for AI-generation tells (broken hands, garbled signage, impossible lighting, over-smooth skin) and returns visual signals alongside the verdict.
• Honest uncertainty. Every verdict comes with a confidence score and a short usability note — “very short text, low signal”, “low resolution masks fine detail”. No tool can give absolute answers; we don’t pretend otherwise.
• Any language. Pick the response language in the popup. The extension detects the input language automatically and writes the verdict, summary, and claims in the language you chose.

Privacy first

The extension only talks to vsdev.cc. No third-party analytics in the extension bundle, no tracking across pages, and nothing leaves your browser until you explicitly click Check.

• Text you paste is sent once, analysed, and never stored beyond an aggregate usage counter.
• Images are re-encoded on your device (max 1024px on the longest side) before upload — the source URL is retained only for the current request, never persisted.
• Your Google sign-in gives us only your email — used purely to key your quota and subscription.
• Press Esc during image picking to cancel at any time.

How it works

1. Sign in once with Google — the same account you use for Drafto and FastNews.
2. Paste text or click “Check image on page”. For images, click the one you want; the extension redraws it into a canvas and uploads the bytes.
3. Your backend calls OpenAI with a strict-JSON prompt tuned for authenticity checks, then returns a sanitised verdict.
4. The popup shows verdict + confidence + summary + claims/signals. Clear and run another check any time.

Plans