Google's A2A protocol and a Jakarta EE server implementation
summary
Google A2A is a prompt‑forwarding RPC layer where one agent sends a user prompt and the receiving agent decides how to execute tools rather than relying on predefined function schemas A2A sends prompts, not schemas contrasted with MCP which requires explicit input/output definitions MCP uses function schemas. Payload follows MCP OPC 2.0 style with status updates for long jobs and artifacts like images. A Jakarta EE server implementation and official A2A SDK provide annotations that expose agents as endpoints and a client that discovers agent cards via a base URI.
AI-written summary · about
related talks
47mAgent Client Protocol for Java Developers
19mgRPC as the Native Transport for Agent Protocols (A2A and MCP)
21mAre APIs becoming obsolete? Think MCP, ACP, A2A, AGNTCY etc.
32mFrom Monolith to AI Agent Modernizing Java Systems with MCP
11mDon't let AI agents push your buttons - use webMCP instead!
24mUnlocking AI Agentic UX with Google’s AP2 Standard and the x402 Protocol
46mAI Tools for Jakarta EE
27mBuilding AI Agents with Spring & MCP