Was this page helpful?
Connect ChatGPT, Claude, Gemini, and custom runtimes to Attribution.ai through the MCP JSON-RPC transport.
Use OAuth + PKCE when your client supports it. Use API keys for automation or clients that do not complete OAuth flows.
https://attribution.ai/api/mcp/rpc
Send JSON-RPC requests with tools/list and tools/call methods to execute workspace-scoped measurement tools.
Use OAuth by default for interactive connector setup.
Discovery metadata:
/.well-known/oauth-authorization-server/.well-known/oauth-protected-resourceOAuth endpoints:
POST /api/mcp/authorizePOST /api/mcp/tokenPOST /api/mcp/registerService automation
https://attribution.ai/api/mcp/rpc.https://attribution.ai/api/mcp/rpc.https://attribution.ai/api/mcp/rpc.For internal copilots, autonomous analysts, or CI jobs, use the same endpoint and auth model as interactive connectors. For terminal-first pipelines, pair this with the CLI.
If a client cannot complete OAuth, use a bearer API key generated from Connect Your AI.
curl -s -X POST https://attribution.ai/api/mcp/rpc \
-H "Content-Type: application/json" \
-H "Authorization: Bearer atai_your_key_here" \
-d '{"jsonrpc":"2.0","method":"tools/list","id":1}'initialize, tools/list, and tools/call succeed.WWW-Authenticate resource metadata.