Linux Foundation's A2A Protocol Hits v1.0 Letting AI Agents Talk Across Any Cloud
Google's Agent2Agent protocol hits v1.0 under Linux Foundation governance, with signed Agent Cards, multi-tenancy, and cross-framework interoperability now stable.
Takeaways
- A2A protocol hits v1.0, first production-ready release of the open agent interoperability standard.
- Governed by Linux Foundation with TSC from AWS, Cisco, Google, IBM, Microsoft, Salesforce, SAP, ServiceNow.
- Adds signed Agent Cards, multi-tenancy, version negotiation, and modernized security for enterprise deployments.
- Built on JSON-RPC 2.0 over HTTP with SSE streaming and webhooks; also supports gRPC bindings.
- Complements MCP: A2A handles agent-to-agent, MCP handles agent-to-tool. See the repo.
- Official SDKs available for Python, Go, JavaScript, Java, .NET, and Rust with backward-compatible migration path from v0.3.
The Agent2Agent (A2A) protocol just hit its first stable release, giving multi-agent systems a common wire format that spans frameworks, clouds, and vendors. The repo now sits at over 25,000 stars, and v1.0 marks the moment the spec is officially considered production-ready by the community shepherding it.
A2A is an open-source communication standard that enables autonomous AI agents to discover, authenticate, and interact with each other regardless of their underlying implementation or hosting platform. Released under Apache-2.0 licensing and now governed by the Linux Foundation, it provides the interoperability layer that multi-agent systems require to operate across organizational and vendor boundaries. If MCP is the standard for how one agent talks to its tools, A2A is the standard for how agents talk to each other.
The problem it actually solves
Anyone wiring up more than one agent has hit this wall: a LangGraph orchestrator needs to hand a subtask to a CrewAI specialist, which then needs to consult a bespoke agent running behind a partner's firewall. Today that means bespoke glue code for every pairing. A2A addresses this: agents built with different frameworks, deployed on different clouds, and operated by different organizations need to collaborate on complex workflows without requiring custom integration code for each connection.
With A2A, agents can discover each other's capabilities, negotiate interaction modalities (text, forms, media), securely collaborate on long-running tasks, and operate without exposing their internal state, memory, or tools. That last property is what the spec calls opacity, and it is why enterprises care: your agent can cooperate with a partner's agent without leaking its prompt graph, memory store, or tool credentials.
How the wire protocol actually works
The design is deliberately boring in the best way. A2A achieves interoperability through standardized discovery using Agent Cards, structured task management with defined lifecycle states, and enterprise-grade security including OAuth 2.0, API Keys, and mTLS support, all built on familiar web technologies (HTTP/HTTPS, JSON-RPC 2.0).