Devtoolwatch BoardGet it every Monday

A2A Protocol review

A2A Protocol is an open protocol that lets opaque AI agents built on different frameworks discover each other, communicate, and collaborate.

Updated Nov 1, 2026. Review written Oct 25, 2026 from the project's own sources; usage data refreshed every Monday.

TL;DR

A2A Protocol (Agent2Agent) is an open standard, originated by Google and now hosted by the Linux Foundation, for communication between AI agents. It is for developers who need agents from different frameworks and vendors to work together. Agents collaborate as agents, not just as tools, without exposing internal state, memory, or tools.

A2A Protocol at a glance

Category
Agent frameworks
Maintained by
Agent2Agent (A2A) Project (organization, @a2aproject on GitHub)
License
Apache 2.0
Language
Shell, Python
Repository
github.com/a2aproject/A2A
Website
a2a-protocol.org
PyPI package
a2a-sdk 1.2.3
Latest release
a2a-sdk 1.2.3 on PyPI, Oct 19, 2026
First published
Mar 25, 2025

What A2A Protocol does

  • A2A Protocol lets agents discover each other's capabilities through Agent Cards that list capabilities and connection info.
  • A2A Protocol lets agents negotiate interaction modalities such as text, forms, and media.
  • A2A Protocol supports synchronous request/response, streaming over SSE, and asynchronous push notifications.
  • A2A Protocol exchanges text, files, and structured JSON data between agents.
  • A2A Protocol supports secure collaboration on long-running tasks without agents sharing internal memory, proprietary logic, or tool implementations.
  • A2A Protocol security schemes include mTLS and OAuth2 metadata, and skills can specify their own security.
  • A2A Protocol Agent Cards can carry signatures and announce multiple transports.

How it works

A2A Protocol uses JSON-RPC 2.0 over HTTP(S). Each agent publishes an Agent Card at a well-known URI, agent-card.json, describing its capabilities and connection details. Clients read the card, then send tasks and receive responses, streams, or push notifications. Official SDKs exist for Python, Go, JavaScript, Java, .NET, and Rust.

How to install and start using A2A Protocol

Python SDK with pip

pip install a2a-sdk

JavaScript SDK with npm

npm install @a2a-js/sdk

Go SDK

go get github.com/a2aproject/a2a-go

Requirements

  • The A2A Python SDK (a2a-sdk) requires Python 3.10 or later.

First steps

  1. Read the A2A Protocol documentation site for the overview, specification, tutorials, and guides.
  2. Review the A2A Protocol Specification to understand Agent Cards, tasks, and transports.
  3. Install the SDK for your language, such as the Python, Go, JS, Java, .NET, or Rust SDK.
  4. Run the official a2a-samples repository to see A2A agents in action.

Who A2A Protocol is for

  • Developers connecting agents built on different frameworks, such as LangGraph, CrewAI, Google ADK, or Genkit.
  • Teams building sequential or hierarchical multi-agent workflows across separate servers or companies.
  • Organizations that need agents to collaborate without revealing proprietary logic or internal memory.
  • Developers exposing existing agents as A2A servers or writing A2A clients.

Limitations to know

  • A2A Protocol is a specification. You implement it with an SDK or your own code rather than running a standalone tool.
  • A2A Protocol has shipped breaking spec changes across versions, including v0.2.5, v0.3.0, and v1.0.0.
  • A2A Protocol features such as a QuerySkill() method and dynamic UX negotiation within a task are still planned, not finished.
  • A2A Protocol lists streaming reliability and push notification improvements as future work.

Is A2A Protocol growing? Usage data

In the week of Oct 25–31, 2026, A2A Protocol ranked #211 on the Devtoolwatch Board and #22 in Agent frameworks, with a momentum score of 43 out of 100. Its main package had 2.9M downloads that week, +6% against its own 4-week average.

New GitHub stars per week for A2A Protocol12 weeks from Aug 9, 2026: 114, 113, 98, 107, 112, 118, 88, 74, 118, 98, 104, 89 new stars, an average of 103 a week.Aug 9week of Oct 25
New GitHub stars per week for the last 12 weeks, from GitHub's stargazer history (today's stargazers by the week they starred).
Show the weekly numbers
New GitHub stars per week for A2A Protocol
Week ofNew stars
Aug 9, 2026114
Aug 16, 2026113
Aug 23, 202698
Aug 30, 2026107
Sep 6, 2026112
Sep 13, 2026118
Sep 20, 202688
Sep 27, 202674
Oct 4, 2026118
Oct 11, 202698
Oct 18, 2026104
Oct 25, 202689

The Board ranks tools by four signals each week: new stars, new stars for the repo's size, added downloads and download growth beyond the market. How the ranking works.

A2A Protocol alternatives

Other AI agent frameworks on the Board in the week of Oct 25–31, 2026, by momentum:

Agent frameworks ranked by momentum, the week of Oct 25–31, 2026
RankToolStarsDownloads/wk
#1Harness²Harness² is a Python framework from Google Research for recursive agent harnessing, refining an agent's prompts, tools, memory and…21.7k–
#2HyperFramesHyperFrames is an open-source framework that turns HTML, CSS, media, and seekable animations into deterministic MP4 videos, built for AI…76.6k1.5M
#3AgentScopeAgentScope is a Python agent framework with an SDK for building ReAct agents and an agent service backend with a web UI.35k51.4k
#4AG-UIAG-UI is an open, event-based protocol that standardizes how AI agents connect to user-facing frontend applications.16.9k4.6M
#5LangGraphLangGraph is a low-level orchestration framework and runtime for building, managing, and deploying long-running, stateful agents in…44.2k18.5M
#6BifrostOpen-source, high-performance AI gateway that works as your LLM gateway, MCP gateway and agent gateway. One OpenAI-compatible API for…9.3k–

All AI agent frameworks, ranked

Recent A2A Protocol releases

A2A Protocol FAQ

Is A2A Protocol free and open source?

Yes. A2A Protocol's source code is on GitHub (a2aproject/A2A) under the Apache 2.0 license, so you can use it for free.

How do I install A2A Protocol?

Run pip install a2a-sdk. The a2a-sdk package needs Python 3.10 or newer. The install section above has the other options from the project's README.

Is A2A Protocol still maintained?

Yes. The a2aproject/A2A repository was last updated on Oct 21, 2026, and the latest release, a2a-sdk 1.2.3 on PyPI, came out on Oct 19, 2026.

How popular is A2A Protocol?

A2A Protocol has 26.4k GitHub stars as of Nov 1, 2026. It gained 89 new stars in the week of Oct 25–31, 2026, against a 12-week average of 103 a week. Its main package had 2.9M downloads that week, +6% against its own 4-week average. That made it #22 among AI agent frameworks on the Devtoolwatch Board.

A2A Protocol vs LangGraph: how do they compare?

A2A Protocol and LangGraph are both tracked by Devtoolwatch as AI agent frameworks. A2A Protocol has 26.4k GitHub stars and LangGraph has 44.2k. In the week of Oct 25–31, 2026, A2A Protocol's main package had 2.9M downloads and LangGraph's had 18.5M. LangGraph is a low-level orchestration framework and runtime for building, managing, and deploying long-running, stateful agents in Python or JavaScript. LangGraph review.

What are the best alternatives to A2A Protocol?

Other AI agent frameworks on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: Harness² (#1), HyperFrames (#2), AgentScope (#3), AG-UI (#4), LangGraph (#5). The alternatives table above has their numbers.

What is the A2A Protocol in AI?

A2A (Agent2Agent) is an open protocol that lets generative AI agents built on different frameworks, by different companies, and on separate servers communicate and collaborate as agents. Google originated it, and the Linux Foundation now hosts it.

How does the A2A Protocol work?

A2A uses JSON-RPC 2.0 over HTTP(S). Agents advertise capabilities in Agent Cards, then exchange text, files, and JSON through synchronous requests, SSE streaming, or asynchronous push notifications.

How does A2A relate to MCP?

A2A complements MCP. A2A focuses on agents collaborating with each other as agents, not just as tools. The official DeepLearning.AI course covers using A2A and MCP together.

Does LangGraph support the A2A Protocol?

Yes. Agents built with LangGraph, Google ADK, or BeeAI can be exposed as A2A servers. The homepage also lists CrewAI and Genkit among the compatible frameworks.

Do agents have to share their internal state with A2A?

No. A2A lets agents operate without exposing internal state, memory, proprietary logic, or tool implementations. This opacity is a stated design goal of the protocol.

Is the A2A Protocol production ready?

A2A reached version 1.0.0 in March 2026, followed by 1.0.1. The project says it is designed with security, authentication, and observability in mind.

Sources

  1. a2aproject/A2A on GitHub
  2. A2A Protocol README
  3. a2a-sdk on PyPI
  4. A2A Protocol website
  5. Release v1.0.1 (May 28, 2026)
  6. Release v1.0.0 (Mar 12, 2026)
  7. Release v0.3.0 (Jul 30, 2025)
  8. Devtoolwatch Board, the week of Oct 25–31, 2026

A2A Protocol README badge

Maintainers can show the tool's weekly Board rank in their README:

Devtoolwatch badge for A2A Protocol

[![Devtoolwatch](https://index.devtoolwatch.com/badge/a2a.svg)](https://index.devtoolwatch.com/tools/a2a/)

How this page is made: Devtoolwatch writes each review from the project's README, releases, package metadata and website, with an AI model, then checks every claim against the source it cites and drops what the sources don't support. The numbers come from Devtoolwatch's daily GitHub snapshots, GitHub's stargazer history, and the npm and PyPI download APIs. Spotted an error? Email [email protected].