Free public utility · No login required
AI Watermark Detector
& Remover
Detect and strip invisible watermarks, C2PA manifests, cryptographic signatures, hidden Unicode, and AI provenance metadata embedded by Claude, ChatGPT, Gemini, and others — while preserving your original content exactly.
Input
Click to upload or drag & dropDOCX, PDF, RTF, ODT, images, code, text · max 20MB
OR PASTE TEXT
0 chars · 0 words
Metadata, Unicode, C2PA
What this tool does
- CONFIRMED detection: C2PA manifests (Anthropic, OpenAI, Google, any provider)
- CONFIRMED detection: hidden Unicode (zero-width, tag chars, BOM)
- CONFIRMED detection: AI metadata in DOCX, PDF, images, SVG
- CONFIRMED detection: cryptographic signatures binding provenance
- CONFIRMED detection: LSB steganography in PNG/BMP (deep scan)
- CONFIRMED detection: hidden Unicode + trailing bytes in code files
- LIKELIHOOD gauge: if no watermark is found, estimates % chance of AI processing
- Strips all CONFIRMED watermarks, preserving visible content
- Re-scans to verify removal — binary: CLEAN or PARTIAL
- Generates SHA-256-hashed evidence report (JSON/MD/HTML)
What this tool is NOT
- ✕ It is NOT an AI text detector (we don't care if AI wrote the text)
- ✕ It is NOT a plagiarism checker
- ✕ It is NOT an academic integrity tool
- ✕ It does NOT judge whether AI use was appropriate
- ✕ It does NOT report users to anyone
- ✕ It does NOT log or retain documents
AI companies do not own the text they process. If a human writes 10,000 words and an AI fixes one comma, the human owns all 10,000 words. Invisible watermarks that claim otherwise are an overreach. This tool restores your authorship by removing markers you never consented to.