01
Signal 01Watermark detection
The scan reports a watermark-detected percentage for the submitted passage so you can see the result immediately and act on it.
Forensic AI watermark scanner
Forensic AI watermarkscanner for generated text.
Run a technical AI watermark scan on up to 500 words. Flag watermark signals, then remove the watermark with StealthGPT. One free detection. No signup.
- Technical watermark scan
- One free detection
- Then remove with StealthGPT
MODE / TEXT FORENSICSENDPOINT / DETECTLIMIT / 500 WORDS
Live detector01 / 01 free scan
Detect AI watermarksOne free scanNo signupThen remove with StealthGPT
Scanner model
A forensic pass built to catch AI watermark signals.
This scanner checks generated text for watermark signals, then routes you to StealthGPT to remove them. It is a detector and checker, not an official model-provider verifier.
02
Signal 02Surface diagnostics
Rule-based cues flag repetition, stock transitions, and sentence-length regularity so you can inspect the prose while you decide to remove the watermark.
03
Signal 03Removal path
After the free scan, open StealthGPT’s watermark remover to rewrite the draft and strip watermark patterns from the wording.
Protocol / AWF-01
From specimen to watermark removal in three steps.
Paste text, run one free forensic detection, then continue in StealthGPT to remove the watermark. The API key stays on the server.
- 01
Validate the specimen
Whitespace-only, oversized, cross-origin, and repeat free-scan requests are rejected before they can spend detector credits.
- 02
Run the watermark scan
The server posts the passage to StealthGPT’s detector endpoint with caching disabled and never exposes the API token.
- 03
Remove the watermark
Read the watermark-detected result, then rewrite the draft in StealthGPT to remove AI watermark patterns from the wording.
Questions
Forensic AI watermark scanner FAQs
01What does this forensic AI watermark scanner do?+
It scans submitted text for AI watermark signals and returns a watermark-detected result. After the free scan, use StealthGPT to remove the watermark from the wording.
02Is this a Unicode or hidden-character scanner?+
No. It analyzes ordinary text through StealthGPT’s detector and separately shows simple linguistic surface cues. It does not claim invisible-character scanning.
03Can I run unlimited forensic scans?+
No. Each browser gets one free detection per detector domain. After that, continue in StealthGPT to remove watermarks with higher limits.
04Does a high watermark-detected score prove authorship?+
No. It flags watermark signals in the passage. It is not a court-ready forensic certificate and does not identify a specific model.
05How do I remove the watermark after the scan?+
Open StealthGPT’s Claude watermark remover from the result popup. StealthGPT rewrites the draft so watermark patterns in the wording are stripped.