What is Langgraph A2a Mcp Example

langgraph-a2a-mcp-example is a repository that demonstrates how to build an A2A Protocol-compatible application using LangGraph with Multi-Channel Protocol (MCP) capabilities.

Use cases

Use cases include building chatbots that require interaction between multiple agents, applications needing persistent state management, and systems that benefit from human oversight in AI operations.

How to use

To use langgraph-a2a-mcp-example, set up the LangGraph application and the LangGraph A2A Adapter. Follow the instructions in the repository’s README for installation and configuration.

Key features

Key features include A2A Protocol support for seamless integration with A2A-compatible clients, Multi-Channel Protocol (MCP) for structured communication between AI systems, stateful conversation support, and human-in-the-loop capabilities.

Where to use

langgraph-a2a-mcp-example can be used in various fields such as AI development, conversational agents, and multi-agent systems requiring structured communication.

LangGraph A2A MCP Example

This repository demonstrates how to build an A2A Protocol-compatible application using LangGraph with Multi-Channel Protocol (MCP) capabilities.

Video Overview

Architecture

graph TD
    A2AClient[A2A Client] -->|A2A Protocol| A2AAdapter[LangGraph A2A Adapter]
    A2AAdapter -->|LangGraph Server API| Graph[LangGraph Agent Graph]
    Graph -->|MCP| Tavily[Tavily Search]
    
    subgraph "This Repository"
        Graph
        Tavily
    end
    
    style A2AClient fill:#e1e8ed,stroke:#333,stroke-width:1.5px
    style A2AAdapter fill:#9aadc2,stroke:#333,stroke-width:1.5px
    style Graph fill:#c2dcf2,stroke:#333,stroke-width:1.5px
    style Tavily fill:#d3e0ea,stroke:#333,stroke-width:1.5px

Overview

LangGraph is a library for building stateful, multi-actor applications with LLMs. This example shows how to create a LangGraph application that is compatible with the A2A (Agent-to-Agent) Protocol, enabling it to communicate with any A2A-compatible client.

Key features:

• A2A Protocol Support: Seamless integration with A2A-compatible clients through the adapter
• Multi-Channel Protocol (MCP): Support for structured communication between AI systems using LangChain MCP Adapters
• Stateful Conversation: Built-in support for persistent state, checkpoints, and multi-step interactions
• Human-in-the-Loop: Capability for both autonomous operation and human collaboration

Setup Requirements

You’ll need to set up two components:

1. This LangGraph application - The actual agent implementation
2. LangGraph A2A Adapter - Translates between A2A protocol and LangGraph API (GitHub repo)

1. Setting up the LangGraph Application

1. Clone the repository:

   git clone <repository-url>
   cd langgraph-a2a-mcp-example
   

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

3. Install dependencies:

   pip install -r my_agent/requirements.txt
   

4. Create a .env file with your API keys:

   cp .env.example .env
   

   Then add your API keys for Anthropic, Tavily, and OpenAI.

2. Setting up the A2A Adapter

1. Clone the adapter repository:

   git clone https://github.com/n-sviridenko/langgraph-a2a-adapter.git
   cd langgraph-a2a-adapter
   

2. Follow the installation instructions in the adapter’s README.

3. Create a .env file with the following configuration:

   # LangGraph Connection
   LANGGRAPH_API_URL=http://localhost:2024
   
   # A2A Server Configuration
   A2A_PUBLIC_BASE_URL=http://localhost:8000
   A2A_PORT=8000
   
   # Agent Card Configuration
   AGENT_NAME="Weather Assistant"
   AGENT_DESCRIPTION="An AI assistant that provides weather information, forecasts, and related climate data."
   AGENT_VERSION=1.0.0
   AGENT_SKILLS='[{"id":"weather_info","name":"Weather Information","description":"Get current weather conditions for any location","examples":["What\'s the weather like in New York?","Is it raining in London right now?"]},{"id":"weather_forecast","name":"Weather Forecast","description":"Get weather forecasts for upcoming days","examples":["What\'s the forecast for Tokyo this weekend?","Will it snow in Chicago next week?"]}]'
   

Running the System

1. Start the LangGraph application:

   langgraph dev
   

2. In a separate terminal, start the A2A adapter:

   cd langgraph-a2a-adapter
   python main.py
   

3. Connect any A2A-compatible client to the adapter at http://localhost:8000. You can use the Google A2A Demo Web App for testing.

A2A Integration

The A2A adapter provides:

• Agent discovery through standard A2A agent cards
• Message exchange with assistants
• Task management
• Streaming responses
• Push notifications for task updates

Deployment

In order to deploy this agent to LangGraph Cloud you will want to first fork this repo. After that, you can follow the instructions here to deploy to LangGraph Cloud.

For the A2A Adapter, see the deployment instructions in its repository.