Roadmap

Where OpenAIIS is today.

OpenAIIS is in preview. The protocol is specified and a reference implementation works; the open, independently operated network comes next.

Available now

In the reference implementation.

  • Create and verify signed claims
  • Typed claims for works, identities, consent, authority and lifecycle events
  • Append-only registry with inclusion and consistency proofs
  • Signed checkpoints and independent receipt checks
  • Local registry and read-only verifier services
  • Bitcoin anchoring of checkpoints, with stated assurance levels
Later

What the open network will add.

  • A permissionless network of independently run nodes
  • Hosted registration on the Internet
  • Bitcoin-confirmed proofs available for every record

The reference implementation is a release candidate written in Python. Its release checks are smoke tests, not a full security assessment or network-level adversarial testing.

Read the earlier whitepaper (V1.0, March 2026)? Here is what changed.

The first whitepaper described a blockchain giving every AI agent and piece of software a permanent cryptographic identity. The current documents keep that aim and change several things:

  • Scope. OpenAIIS now covers digital claims generally: people, organisations, creative and research works, software, identities and autonomous systems, not only AI agents.
  • Model. Registering a claim, gathering the related records and deciding are separate steps, and each relying party decides with its own policy.
  • History. The working core is an append-only Merkle registry with signed checkpoints and optional anchoring to Bitcoin.
  • Network and economics. A proof-of-work network remains the intended architecture. Fees for services or network capacity are now allowed.
  • Licensing. Software under Apache 2.0, documents under CC BY 4.0, replacing the earlier MIT licence. See Licensing.
  • Name. OpenAIIS now stands for Open Artificial Intelligence Integrity Sentinel.