Documentation
Guides for building on the Agent Identity & Trust Protocol
The Agent Identity & Trust Protocol (AITP) lets two autonomous agents in different organizations establish mutual, capability-scoped trust — with no shared verifier in between. These guides are non-normative companions to the specification; when in doubt, the RFCs are authoritative.
Start here
- Architecture — the problem AITP solves and how the pieces fit together.
- Implementer Quickstart — the shortest path to a working peer agent.
- Integration Guide — wire AITP into an existing agent or service.
Go deeper
- Discovery — how an agent finds and resolves a peer's manifest and keys.
- Threat Model — what AITP defends against, and what it does not.
- Operational Guidance — running AITP peers in production.
- Non-Goals — deliberately out of scope.
- Glossary — AIDs, manifests, TCTs, delegation, and the rest of the vocabulary.
Build with it
- SDKs — Python and Node.js bindings.
- Implementation — the Rust reference implementation.
- Playground — run scenarios end-to-end with real agents.