Enable the kapa.ai MCP endpoint for the OpenFGA docs (like opentelemetry.mcp.kapa.ai) #560
Closed
jensholdgaard
started this conversation in
Ideas
Replies: 2 comments
|
Thanks for the suggestion! We enabled it here https://openfga.mcp.kapa.ai, we'd love the PR! :) |
0 replies
|
It is now discoverable directly from the OpenFGA website through Ask AI widget!
|
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment

Uh oh!
There was an error while loading. Please reload this page.
Hi — I'm building an OpenTelemetry log backend (Ourios) and I'm designing its authorization on OpenFGA: a coarse-tenant + graph-fed visibility model over the OTel GenAI semantic conventions (conversations, agents, users as relationship nodes; agents as first-class principals; MCP tools as FGA objects). Your kapa.ai assistant has been genuinely useful for that — two long design reviews so far, with accurate pointers into the docs and sample stores.
The ask: would you enable kapa's MCP server surface for the OpenFGA docs, the way the OpenTelemetry project has at
https://opentelemetry.mcp.kapa.ai(a single hosted tool,search_<project>_knowledge_sources)? It's a switch on the kapa side rather than new infrastructure, and it would let coding agents (Claude Code, Cursor, Copilot, …) query the same knowledge base while authoring models — the natural complement to the new openfga/agent-skills skill: the skill carries the rules, the MCP answers the questions the rules don't.Concretely, in our workflow the OTel docs MCP is wired as an MCP server in the agent, and every semantic-convention or SDK question gets grounded in the docs before code is written; we'd do the same for OpenFGA modelling and SDK questions. If it's enabled, I'm happy to send a small PR adding it to the AI Tooling section of this repo's README next to Agent Skills / DeepWiki / Context7.
Thanks for the project — the agents-as-principals and MCP-authorization guides are exactly the shape we needed.
All reactions