Sign content before AI trains on it.
Three-layer cryptographic timestamps that prove content existed at a specific moment. Useful for opt-out registries, licensing claims, training-cutoff documentation, and rights enforcement against unauthorized scraping.
Why training data provenance matters
Opt-out enforcement
Frontier model providers offer opt-out mechanisms. Cryptographic timestamps prove your content existed before the opt-out and was protected against later training runs.
Licensing claims
If you license content for AI training, signed receipts establish the exact corpus a licensee received and when, separating licensed use from unlicensed scraping.
Training cutoff documentation
When a dispute arises about whether content was in a model's training data, the signed timestamp shows whether the content existed before or after the training cutoff date.
FRE 902 admissibility
Receipts are designed for self-authenticating digital evidence under FRE 902(13)/(14). RFC 3161 timestamps from an independent TSA + optional Bitcoin anchor.
Sign content via the same API as AI output
AI Provenance signs both AI-generated output and pre-AI content. Same SDK, same three-layer cryptographic stack, same verification flow. 100 signings/month free, then $0.01/signing with volume discounts.
How the cryptographic stack works →