Call the A2A guide lookup service

Send a short text question and receive public guidance, known limitations, and the best matching page. The service is immediate and stateless.

HTTP+JSON

Send a JSON body to POST https://pdfmdtools.com/a2a/message:send with Content-Type: application/a2a+json. The body contains message.messageId, message.role and a text part.

{"message":{"messageId":"example-1","role":"ROLE_USER","parts":[{"text":"How can I make a Markdown table fit in a PDF?"}]}}

The response contains an agent message with text guidance and a structured site_lookup_result part.

JSON-RPC

Send POST https://pdfmdtools.com/a2a with a JSON-RPC 2.0 SendMessage request.

{"jsonrpc":"2.0","id":1,"method":"SendMessage","params":{"message":{"messageId":"example-1","role":"ROLE_USER","parts":[{"text":"Can this tool read scanned PDFs?"}]}}}

Input and output boundaries

• Input is one text question, up to 1,000 characters.
• Results are deterministic matches against public guidance pages.
• Supported topics include tables, scanned PDFs, columns, local handling, and Markdown-to-PDF workflows.
• There is no file upload, account, authentication flow, agent memory, streaming response, or task queue.

Read the Agent Card before integrating, and use the OpenAPI document for the HTTP+JSON endpoint description.

For the protocol context around these requests, compare A2A, MCP and WebMCP roles.

Next: privacy and capability boundaries · All AI agent guides