-
Notifications
You must be signed in to change notification settings - Fork 44
Open
Description
Context
AI agents building integrations or investigating data flows need to know what API endpoints exist, what they accept, and what they return. API schemas are a key part of organizational context that Meteor doesn't currently capture.
Scope
New extractor(s) for API schema formats:
OpenAPI / Swagger
- Extract endpoints: path, method, parameters, request/response schemas
- Extract service metadata: title, version, description, servers
- Emit as
applicationor newapiasset type with structured schema data - Emit
depends_on/producesrelationships to data assets where inferrable
gRPC / Protobuf
- Extract service definitions: service name, methods, request/response types
- Extract from proto files or gRPC reflection
- Emit relationships to other services and data assets
Design Considerations
- Source can be local files (in a repo), a URL (hosted spec), or a schema registry
- Should work with Stencil schema registry as a source
- Schema data should be structured enough for AI agents to generate API calls
Why
API schemas connect the service layer to the data layer. An AI agent that knows both the database schema and the API schema can reason about the full stack.
References
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels