- Runs it
- Linux Foundation project
- Projects
- 20
- Official
- 11
- Stars checked
- 2026-11-08
- Spec
- official
A2A SDKs, Samples and Tools: 20 open-source projects compared
Find an A2A SDK or tool you can build on today. Compare 20 open-source Agent2Agent projects by GitHub stars, licence, last activity and whether the official a2aproject team maintains them. Official SDKs cover Python, JavaScript, Java, Go, .NET and Rust.
Compare A2A projects
These 20 open-source projects implement or support A2A. 11 are maintained by the official project and the rest by the community. Sort by stars or recent activity, or filter by type. These projects are listed from public GitHub data and are not scanned by AgenticMart.
A2A Inspector
Tool for inspecting and validating A2A agents
A2A CLI
Command-line interface for talking to A2A-compatible agents
A2A TCK
Technology compatibility kit for testing A2A implementations
Stars and activity checked 2026-11-08. These projects are listed from public GitHub data and are not scanned by AgenticMart. Built one? List it free now, paid selling opens soon.
What is A2A?
A2A (Agent2Agent) is an open standard for communication between AI agents. It covers agent-to-agent interaction, while MCP covers connections from agents to tools. Google originally developed it and donated it to the Linux Foundation.
A2A vs MCP: what is the difference
A2A connects agents to other agents, while MCP connects one agent to its own tools, APIs and data. The A2A docs give the example of a mechanic agent that uses MCP to reach a diagnostic scanner and A2A to order a part from a supplier's agent. A2A is aimed at stateful, multi-turn work that goes beyond a single tool call. New to MCP? Read what an MCP server is.
Official A2A SDKs and how to install them
The A2A project publishes official SDKs for Python, JavaScript, Go, .NET, Rust and Java, all implementing the same v1.0 specification. Pick the one for your language, install it with the command below, then start from the official samples.
| Language | Install | Project |
|---|---|---|
| Python | pip install a2a-sdk | View → |
| JavaScript / TypeScript | npm install @a2a-js/sdk | View → |
| Go | go get github.com/a2aproject/a2a-go | View → |
| C# / .NET | dotnet add package A2A | View → |
| Rust | cargo add a2a-lf | View → |
| Java | See the repository README for Maven coordinates | View → |
Install commands quoted from the a2aproject/A2A README and PyPI, checked 12 November 2026; not run for this page.
What an A2A message looks like
A client agent sends a user message to a remote agent over the A2A HTTP+JSON binding, which maps to the SendMessage operation in the A2A 1.0 spec. The remote agent replies with either a Message or a Task object that tracks the work and its artifacts.
POST /message:send HTTP/1.1
Host: agent.example.com
Content-Type: application/a2a+json
{
"message": {
"role": "ROLE_USER",
"parts": [{"text": "What is the weather today?"}],
"messageId": "msg-uuid"
}
}From the official specification, checked 2026-11-12.
A2A questions
What is the A2A protocol?
A2A (Agent2Agent) is an open standard that lets AI agents from different vendors and frameworks talk to each other. Agents use it to learn what another agent can do, hand over work and share results. Each agent keeps its memory, logic and tools private.
A2A vs MCP: what is the difference?
A2A connects agents to other agents, while MCP connects one agent to its own tools, APIs and data. The A2A docs give the example of a mechanic agent that uses MCP to reach a diagnostic scanner and A2A to order a part from a supplier's agent. A2A is aimed at stateful, multi-turn work that goes beyond a single tool call.
Which languages have official A2A SDKs?
As of November 2026, official A2A SDKs exist for Python, JavaScript, Java, C#/.NET, Go and Rust. The specification defines JSON-RPC, gRPC and HTTP+JSON bindings, so an agent can also implement the protocol without an SDK.
How do I get started with A2A?
Start with the Python quickstart tutorial on a2a-protocol.org, then read the key concepts page and the specification. You can install an official SDK, for example pip install a2a-sdk for Python or npm install @a2a-js/sdk for JavaScript. Sample agents are in the a2a-samples repository, which is listed below.
How does an Agent Card work?
An Agent Card is a JSON file that tells other agents who an agent is, where to reach it, what it can do and how to authenticate. The usual location is /.well-known/agent-card.json on the agent's domain, which a client reads with an HTTP GET request. Cards can also come from a curated registry or direct configuration, and sensitive details can be kept in an authenticated extended card.
Who created A2A?
Google originally developed A2A and donated it to the Linux Foundation. A Technical Steering Committee with members from AWS, Cisco, Google, IBM Research, Microsoft, Salesforce, SAP and ServiceNow now maintains it. As of November 2026, the project site also carries a notice that A2A is joining the Agentic AI Foundation.
What are the risks and limits of A2A?
A2A sets baseline rules, but much of the real protection depends on how each team builds its agent. Production traffic must use HTTPS and servers must authenticate every request, while the choice of authorization logic, data protection at rest and audit logging is left to implementers. Remote agents are opaque by design, and the spec does not define a standard registry API.