Labsco
MCP SERVER · OFFICIAL PROJECT

Chaterimo Read-Only Endpoint

by chaterimo

Two endpoints for one product: a full-access one with 19 tools, and a read-only one with 12.

Official source
Summary
A separate read-only URL is a cruder control than scopes and a more reliable one.

Scopes depend on the server enforcing them correctly on every call. A different endpoint that simply does not have the write tools cannot be tricked into writing, whatever the token says. If you are connecting an assistant that only needs to look, use the read-only address even though the scopes would also cover you — one is a policy, the other is an absence. Worth noting on privacy: the vendor states Gemini free-tier conversations may be used differently from paid API traffic.

What it is

Chaterimo publishes its MCP surface twice. The full endpoint carries all 19 tools including the writes; a separate read-only address exposes 12 and cannot write at all.

What you get

The read tools on either endpoint; the write tools only on the full one. Which endpoint you configure is the permission decision.

Requirements

OAuth 2.0 with PKCE, refresh tokens and dynamic client registration, scoped mcp:read and mcp:write. Two pricing schemes: bring your own OpenAI, Anthropic or Gemini key, or use managed keys; there is a free Learner tier.

Setup effort

Paste a URL, then authorize — add the endpoint to your client, then approve the OAuth consent screen