A2A is an interoperability protocol for agent-to-agent communication and task collaboration. It is useful when independent agents or organizations need a shared language for delegation, status, artifacts, and handoffs.
At a glance
- Organization
- A2A Protocol project
- Runtime role
- Agent interoperability protocol
- Category
- Protocols and Contracts
- Official documentation
- Visit official documentation opens in a new tab
Where it fits in the runtime stack
Layer 5: agentic and application runtime. A2A sits at the boundary between agents rather than inside a single model invocation.
Primary runtime role
Use A2A when runtime architecture includes multiple agents that must exchange tasks or results without sharing private implementation details.
Not the same as
A2A is not an agent development kit, not a direct tool-call protocol, and not a replacement for MCP.
Integration notes
- Use agent cards or equivalent discovery metadata to describe agent capabilities and endpoint expectations.
- Keep cross-agent handoffs auditable with request identifiers, state, artifacts, and completion status.
- Separate local tool execution policy from remote-agent collaboration policy.
Questions before production use
- What identity and trust model applies when one agent delegates to another?
- Which task states, artifacts, and errors must be preserved for replay?
- Can agent collaboration continue asynchronously without losing traceability?
Review and deprecation posture
This profile is reviewed as part of the aRuntime.com quarterly resource audit. If the official documentation moves, the project is archived, or the resource changes scope, this page should be updated with a dated status note rather than silently removed.
Sources and further reading
- A2A Protocol opens in a new tab — A2A Protocol project; official documentation; accessed 2026-06-20 UTC.
Last reviewed: .