MCP

Every Flux API ships with a built-in Model Context Protocol server — five read tools and two opt-in write tools, scoped to the same collections and permissions as the REST endpoints. This section covers connecting a client, choosing an auth scheme, what retrieval costs, and what is not built yet.


Where to go

  • Connect Claude — wire up Claude Code, another MCP client, or a hosted connector using a bearer token.
  • Connect ChatGPT and OpenAI — the OpenAI Responses API path, the ChatGPT UI path, and the current public-vs-private limitation.
  • Preparing a Collection — how to describe a collection so an agent picks the right one, which read methods to connect, and which fields to make searchable.
  • Authentication — which of the three schemes to send, and how fast each way of revoking access takes effect.
  • Retrieval & Cost — what the default page size and text truncation mean for what you get back, and why the bill is dominated by re-reads.
  • Limits — the protocol version served, and what isn't implemented yet.

For the concept and the full tool catalog, start at Agent-Native Flux APIs or go straight to the MCP Server reference.

Was this page helpful?