Agent2Agent protocol, now part of the Linux Foundation, is designed to enable interoperability for AI agents to collaborate effectively. It allows agents to discover capabilities, negotiate interaction methods, and work together on complex tasks while maintaining their internal states. The protocol uses JSON-RPC 2.0 over HTTP with real-time communication facilitated by server-sent events. The implementation has gained traction, with over 100 companies already on board. However, there are concerns regarding its overlap with Anthropic's Model Context Protocol, which may complicate its adoption.
The Agent2Agent protocol will be the cornerstone of a wider Agent2Agent project formed by Google, AWS, Cisco, Microsoft, and others, aiming to foster interoperability for AI agents.
Using the Agent2Agent protocol, agents can discover each other's capabilities, negotiate how to interact, and collaborate securely on long-running tasks, focusing on preserving each agent's internal state.
The protocol is based on JSON-RPC 2.0 over HTTP and uses server-sent events for real-time streaming, allowing agents to communicate effectively through 'agent cards'.
According to Google, the Agent2Agent protocol has seen significant adoption with over 100 companies supporting it, although there have been concerns about its overlap with Anthropic's MCP.
Collection
[
|
...
]