Echo documentation
Technical documentation for Echo — the Ontology Layer that turns scattered sources into one queryable knowledge graph.
Echo turns your repositories, documents, work items, databases, and decisions into one typed knowledge graph that humans and agents query through a single API. These docs explain the knowledge model, the sources Echo ingests, the faceted contract every query flows through, and how to integrate it.
They are written for two readers at once: the person evaluating whether Echo fits their organization, and the engineer who needs to integrate it. Nothing here is marketing shorthand — every API shape, source behavior, and auth path is the one the system actually ships.
Where to start
If you are new, read What is Echo? then follow Getting connected. If you are evaluating the system technically, jump to The faceted contract and the Knowledge API.
How these docs are organized
- Start Here — what Echo is, how it models a project, and the shortest path to a working query.
- The Knowledge Model — sources vs providers, refs, and the contract that keeps the API stable as sources grow.
- Sources — one page per source kind: code, graph, docs, company docs, work items, commits & PRs, database.
- The Knowledge API — discover, search, fetch, activity, and the auth model, with request/response shapes you can predict real responses from.
- MCP tool surface — how the MCP server exposes Echo to any agent.
- Recipes — goal-shaped walkthroughs that cross multiple sources.
Use the search button at the top of the sidebar to jump straight to any page by title or heading.