Certaffy

Feature · Pro / Business

Spot paraphrased AI — only when we're confident

A separate paraphrase signal from raw AI probability. We never invent paraphrase certainty when evidence is thin.

How it works

1

Run a full scan

Paraphrase risk appears alongside human and AI-generated labels when enabled.

2

Check confidence

Purple highlights only when the model exceeds our confidence threshold.

3

Discuss with context

Use explanations and revision tips for fair review conversations.

FAQ

No — we flag AI-style paraphrase patterns, not source copying.

Try it on your own writing

Start free — 3 full highlighted reports, then 1 score-only scan/day up to 800 words. No credit card required.

Start free — no credit card required

Free account: 3 full highlighted reports, then 1 score-only scan/day up to 800 words. No account: 3 score summaries/day up to 2,000 words.