Skip to main content

The communication protocol for AI agents

Project description

AgentMesh SDK

The communication protocol for AI agents.

AgentMesh is an open protocol that defines how AI agents find each other, communicate, build trust, and delegate tasks — across any framework, any model, any cloud.

CI License: MIT Python 3.9+


Why AgentMesh?

Millions of AI agents are being built with LangChain, AutoGen, CrewAI, and custom frameworks. They all work in isolation. There is no standard for:

  • How agents find each other
  • How agents verify each other's identity
  • How agents build trust over time
  • How agents communicate across frameworks

AgentMesh is that standard. Think TCP/IP — but for AI agents.


Install

pip install agentmesh

Quickstart

from agentmesh import Agent

# Create two agents with cryptographic identities
alice = Agent.create("alice")
bob = Agent.create("bob")

# Send a signed message
msg = alice.send(to=bob, intent="Summarize the latest news")

# Verify the signature
print(bob.verify(msg))  # True
print(bob.inbox[0].body["intent_text"])  # "Summarize the latest news"

LangChain Integration

from langchain_anthropic import ChatAnthropic
from agentmesh.adapters import LangChainAdapter

llm = ChatAnthropic(model="claude-sonnet-4-6")
agent = LangChainAdapter(llm, name="my-agent")

# Register in the global AgentMesh network
agent.register(capabilities=["summarization", "translation"])

# Invoke and respond
response = agent.invoke("Explain quantum computing in one sentence.")
print(f"Agent {agent.uri}: {response}")

AutoGen Integration

from agentmesh.adapters import AutoGenAdapter

agent = AutoGenAdapter(
    name="my-autogen-agent",
    system_message="You are a helpful research assistant."
)

agent.register(capabilities=["research", "question_answering"])
response = agent.invoke("What is the capital of France?")

Agent Registry

Agents can find each other by capability:

from agentmesh import Agent
from agentmesh.gateway import GatewayClient

gateway = GatewayClient()

# Register your agent
agent = Agent.create("my-agent")
gateway.register(
    uri=agent.uri,
    name="my-agent",
    capabilities=["summarization"]
)

# Find agents by capability
results = gateway.find(capability="summarization")
print(results)  # [{"uri": "agent://...", "name": "...", "trust_score": 0.0}]

Live Gateway

The AgentMesh gateway is live and free to use:

https://agentmesh-gateway.agentmesh-protocol.workers.dev

Endpoints:

  • GET /health — Gateway status
  • POST /v1/registry — Register an agent
  • GET /v1/registry?capability=X — Find agents by capability
  • POST /v1/send — Send a message

Architecture

Your Agent
    |
    | AgentMesh SDK
    |
AgentMesh Gateway (Cloudflare Workers — global, <30ms)
    |
    | Cloudflare KV
    |
Agent Registry (persistent, TTL-based)

Protocol

AgentMesh is built on open RFCs:


Contributing

AgentMesh is open source and MIT licensed. Contributions welcome.

  1. Fork the repo
  2. Create a feature branch
  3. Add tests
  4. Open a pull request

Roadmap

  • RFC-003 Trust Score Verification
  • CrewAI Adapter
  • Agent-to-Agent HTTP Transport
  • Stripe Usage-Based Billing for Enterprise

Built with Python, Cloudflare Workers, and Ed25519 cryptography.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

agentmesh_protocol-0.1.0.tar.gz (9.4 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

agentmesh_protocol-0.1.0-py3-none-any.whl (11.0 kB view details)

Uploaded Python 3

File details

Details for the file agentmesh_protocol-0.1.0.tar.gz.

File metadata

  • Download URL: agentmesh_protocol-0.1.0.tar.gz
  • Upload date:
  • Size: 9.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for agentmesh_protocol-0.1.0.tar.gz
Algorithm Hash digest
SHA256 22bf6503ebf19ba33c98694b94a9e1a1b3c25496896e8b4aa4e9a60ca50080b8
MD5 bbb1dd1d0457d5a5941ceeab8e2558ef
BLAKE2b-256 5064df28119f63cd5a982a8d3bb543742280a2513e21c0f8489345dfed1d145a

See more details on using hashes here.

Provenance

The following attestation bundles were made for agentmesh_protocol-0.1.0.tar.gz:

Publisher: publish.yml on agentmesh-protocol/agentmesh-sdk

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file agentmesh_protocol-0.1.0-py3-none-any.whl.

File metadata

File hashes

Hashes for agentmesh_protocol-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 98cdfbcb4a48cdbc7c0789d6cf413300f4f98f76c5442e35504738365d7ee732
MD5 3211602528135227a6643a78d995aab9
BLAKE2b-256 a68d928a2461d366980584d43183a1646917b8345088c784414fef97b2331f92

See more details on using hashes here.

Provenance

The following attestation bundles were made for agentmesh_protocol-0.1.0-py3-none-any.whl:

Publisher: publish.yml on agentmesh-protocol/agentmesh-sdk

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page