Every AI action, traced to someone accountable.
OpenAIIS is an open protocol that shows who made a digital claim, who authorised it, who answers for it, and whether it can still be trusted. Anyone can verify it, without relying on a single platform.
- Signed claims
- Verifiable history
- Anchored to Bitcoin
release-bot wants to publish billing-app 2.4.0
- Signed by the expected agent agent:companyx/release-bot
- Agent bound to its operator Company X
- Responsibility accepted Head of Engineering, Company X
- Publishing authority in force Delegated for billing-app releases
- Release matches what was signed SHA-256 digest verified
- Security evidence meets policy Scan report from a trusted issuer
- No revocation on record Checked against shared history
The problem
A valid signature is not permission.
AI agents now write code, publish software and act for companies at machine speed. Today's tools can tell you that a key signed something. They can't tell you whether that key was allowed to, or whether it was stolen an hour ago.
Stolen keys still look valid
A compromised key keeps passing signature checks until every system hears about it. Most never do.
Agents act without a paper trail
An agent can publish, buy or deploy with no portable record of who runs it or who answers for what it did.
Trust is stuck in silos
Every platform keeps its own records. Evidence of a breach at one rarely reaches the next.
How it works
Declare. Record. Decide.
OpenAIIS keeps what was claimed separate from whether it is allowed, so trust can change when the facts do.
-
01
Declare
The agent, the company that runs it and the people responsible each sign what they stand behind: who operates it, who accepts responsibility, and what it may do.
-
02
Record
Every signed claim joins an append-only history that anyone can audit and no one can quietly rewrite. Checkpoints can be anchored to Bitcoin.
-
03
Decide
When the agent acts, any service checks the full chain against its own policy and gets a clear answer: allow, deny or send for review.
And when something goes wrong, the company publishes one signed revocation. Every service that checks the shared history can refuse what the stolen key signs next.
The full walkthroughUse cases
One protocol for anything that needs to be trusted.
AI is the reason OpenAIIS exists, but the same questions apply to every digital claim, whether a person, a company, a program or an agent made it.
AI agents
Bind every agent to the company that runs it and a person who answers for it, and limit what it may do on their behalf.
Software releases
Prove a release came from an authorised publisher, and stop trusting it everywhere the moment its signing key leaks.
Security response
Share signed compromise reports that other organisations can act on automatically, under their own rules.
Creators and researchers
Create independent, timestamped evidence that your work existed, without publishing the work itself.
Media integrity
Show where a photo, article or AI-generated work came from, and what has happened to it since.
Identity and credentials
Issue attestations anyone can check independently, and withdraw them when they no longer apply.
Why OpenAIIS
Trust that doesn't depend on trusting us.
Built from proven cryptography and transparency techniques, arranged so that no single organisation controls the record.
Revocation that travels
Report a stolen key or a bad release once. Every service that checks the shared history can see it and stop.
Your policy decides
OpenAIIS never tells anyone whom to trust. Each service applies its own trusted keys and its own rules.
Private by design
Only fingerprints of your content are recorded. The files themselves stay with you.
Owned by no one
Designed to run on independently operated infrastructure, so trust never hinges on a single company.
History you can prove
Inclusion and consistency proofs show that a record is there and that the history around it was not rewritten.
Works with what you have
Built to sit alongside existing certificates, identity systems and transparency logs, not to replace them.
Already on Bitcoin
Our history, sealed where no one can change it.
On 25 October 2026 a signed OpenAIIS registry checkpoint was confirmed on the Bitcoin blockchain. You can check the block headers yourself, right in your browser.
Verify it yourselfOpen by default
Read it, check it, build on it.
Why OpenAIIS exists
How provenance, governance, responsibility and safety fit into one verifiable system.
Read ProtocolThe exact specification
Formats, signatures, verification rules and conformance, for anyone building an implementation.
Read RoadmapWhere we are today
OpenAIIS is in preview. See what works now, and what comes before the open network goes live.
See roadmapMake every AI action accountable.
Start with the whitepaper, or go straight to the specification.