AIFED is an open protocol exploration for decentralized discovery, trust, and workload federation between sovereign AI operators (e.g., national clouds, regulated providers, and jurisdiction-bound AI fabrics).
This repository publishes protocol ideas in the open and tracks a path toward interoperable reference implementations.
- Draft protocol work
- Not production-ready
- Open to feedback, critique, and design contributions
Current AI platform interoperability is mostly centralized and vendor-shaped. AIFED explores a different model where sovereign operators can federate without a mandatory global authority while preserving policy boundaries and verifiable execution.
- Discovery of sovereign AI nodes
- Cryptographic verification of directory and execution metadata
- Capability and policy advertisement
- Workload submission and verifiable receipts
- Artifact identity using content-addressed references
- Global billing/procurement standards
- Mandatory single registry
- Uniform legal/policy regimes across participants
- Geopolitical governance decisions
docs/spec/project-charter.md: project goals, strategic intent, and core object modeldocs/spec/aifed-1.md: current draft specdocs/spec/changes.md: spec changelogdocs/spec/threat-model.md: threat assumptions and security analysisdocs/spec/interop-profile.md: normative interoperability profiledocs/releases/v0.1.0-draft.md: initial release notes draftdocs/rfcs/: design proposals under discussionreference/openapi.yaml: minimal API surface sketchreference/schemas/: JSON schema stubs for key objectsexamples/: sample payloads and flows
See CONTRIBUTING.md. For security issues, use SECURITY.md.
Licensed under Apache-2.0. See LICENSE and attribution notices in NOTICE.