All docs

MCP Server

Tools

Clients

Cuelara's tools are available as a remote MCP server (Model Context Protocol), so any MCP-compatible AI client — Claude Desktop, Claude Code, Cursor, Copilot, Gemini CLI, and more — can call them directly, without opening the website.

The server is stateless: every request is self-contained, over plain JSON-RPC 2.0 at a single URL, and every tool below is available the moment you connect it.

Endpoint

https://cuelara.com/api/mcp

Fastest setup — Claude Code

bash
claude mcp add --transport http cuelara-token-optimizer https://cuelara.com/api/mcp

For every other client, pick it from the list below for exact config.

Authenticating as yourself

Without a token, calls are anonymous — rate-limited by IP, same as a visitor to the website, and the output gets a "by Cuelara.com" line appended.

Generate a personal token from /dashboard/mcp and pass it as a bearer token — each client page below shows exactly where it goes. Signed-in calls use your own account's daily limit, and paid-plan output skips the attribution line. Free-plan accounts still get the attribution, same as anonymous calls.

Call it directly

No SDK needed — it's plain JSON-RPC 2.0 over HTTP. For example, to list the available tools:

bash
curl -X POST https://cuelara.com/api/mcp \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

Add -H "Authorization: Bearer YOUR_TOKEN_HERE" to any call to authenticate as yourself instead of anonymously. See each tool's page above for its exact arguments and a ready-to-run tools/call example.

Limits

Anonymous calls share the same free daily limit as the matching web tool, keyed by IP. Signed-in calls (via a personal token from /dashboard/mcp) use your own account's plan limits instead.

Integrating from your own code rather than an AI assistant? The same tools are also available as a plain REST API using the same bearer token.