Your content is yours.
AIWaterMarkCheck is engineered from the ground up for private, one-off forensic checks. We inspect your submission in transient memory, return the analysis, and immediately purge the data.
No Disk Storage
Uploaded images, audio clips, and pasted text strings are processed entirely in memory. No files or media buffers are ever written to disk or long-term databases.
Never Used For Training
Your content is never stored, indexed, or pooled to train machine learning models or classifiers. What you check remains strictly confidential.
Zero Personal Tracking
No cookies, tracking pixels, or cross-site fingerprinting scripts. We do not require accounts, logins, credit cards, or personal identifiers.
What we process during a check
- Images & Audio: Media files are streamed to temporary memory to parse C2PA metadata, extract EXIF/XMP tags, and evaluate frequency signatures. Once the analysis completes, the memory buffer is cleared.
- Text Submissions: Pasted text is analyzed for invisible Unicode characters and token distributions in real-time, then discarded.
- Website Scans: We fetch public HTML directly from the target server. Only the public page source is inspected; no credentials or private cookies are stored.
Third-Party Provider Verification
When optional OpenAI provenance verification is selected, the file is checked against OpenAI’s official provenance endpoint. That interaction is governed by OpenAI’s enterprise provenance terms, which also guarantee zero retention for provenance checks.
Operational Logging
Our server logs record strictly operational metrics: timestamp, request duration, HTTP status code, and coarse error codes. We do not log submitted filenames, text content, extracted metadata, or IP addresses.
Local Browser Actions
Exporting a report, copying detected characters, or downloading a cleaned file are client-side operations executed in your browser. No exported reports are uploaded back to our servers.