AgentKit and LangGraph integrations for policy-gated ODEI actions.
This repository explores how orchestration frameworks can route actions through ODEI before execution.
Frameworks can coordinate agent loops. They do not automatically provide:
- durable world state
- policy-aware execution
- structural memory
- pre-execution guardrails
ODEI fills that gap by acting as the governed state and policy layer behind the orchestrator.
- AgentKit integrations
- LangGraph integrations
- guard-before-execute patterns
- world model aware tool routing
- api.odei.ai for production APIs
- mcp-odei for MCP integrations
- research for the architectural rationale
Repository-specific. See local files.