digitalSourceType marking (trainedAlgorithmicMedia)
Generated outputs carry a standards-based, machine-readable AI label, the marking pattern EU AI Act Article 50 calls for.
C2PA for AI Companies
How AI companies use the C2PA standard to mark generated outputs machine-readably, document licensed content provenance, and prepare for EU AI Act Article 50.
The problem
Why AI companies need provenance
Regulators expect machine-readable marking of AI-generated content, with EU AI Act Article 50 applying from August 2, 2026 for new systems. At the same time, training-data lawsuits are turning undocumented ingestion into a balance-sheet risk. Both problems reduce to provenance: marking what you generate and verifying what you ingest.
What you get
What C2PA gives AI companies
C2PA text manifests
LLM text outputs can be marked alongside images and audio, closing the gap most C2PA tooling leaves open.
Verification at ingestion
Licensed content can be verified for origin and integrity as it enters your pipeline, creating a documented record of good-faith sourcing. Verification is free.
Standards-based proof
The official C2PA Conforming Products List records Encypher as a conformant generator product at assurance Level 1, with generation and validation coverage across 20 media types (conformance dated May 26, 2026). Encypher's claim-signing and timestamp authorities appear on the official C2PA Trust Lists. That matters when regulators and enterprise customers review your marking pipeline.
Go deeper
See it on your own content
Generated outputs are marked machine-readably and ingested content is verifiable.