Reference integrations for the ODEI API, MCP, and world model workflows.
This repository contains small, practical examples showing how to connect external agents and applications to ODEI.
- guardrail checks before execution
- world model queries from Python and TypeScript
- MCP integration patterns
- agent workflows that rely on structured memory instead of flat transcripts
- odei.ai for the research surface
- api.odei.ai for production APIs and docs
- validate an action before a tool call
- query active objectives or constraints from the world model
- connect an MCP client to governed retrieval
- prototype ODEI-aware agents in Python or TypeScript
# Python examples
pip install -r python/requirements.txt
# TypeScript examples
cd typescript
npm installSee the example files in this repository for concrete request shapes and setup patterns.
| Repository | Role |
|---|---|
| mcp-odei | MCP server |
| web | Docs and public API runtime |
| memory | Memory architecture |
MIT