mcp-a2a

# Model Context Protocol (MCP) & Agent-to-Agent Protocol (A2A) This repository contains example notebooks for the 2025 course [Agenti Intelligenti e Machine Learning (AiTHO)](https://web.dmi.unict.it/it/corsi/l-31/agenti-intelligenti-e-machine-learning-aitho), focusing on MCP and A2A. ## Tech Stack - **Python** - **[LangChain](https://github.com/langchain-ai/langchain)** – A framework for building AI-based conversational chains - **[LangGraph](https://github.com/langchain-ai/langgraph)** – A framework for building AI agent workflows - **[Streamlit](https://streamlit.io/)** - A fast and intuitive framework for building interactive web apps directly in Python, commonly used for AI and data science applications - **[Model Context Protocol (MCP)](https://modelcontextprotocol.io/introduction)** – A standardized interface for defining, sharing, and exchanging model context across different systems and agents - **[Agent-to-Agent Protocol (A2A)](https://google.github.io/A2A/)** - A communication protocol enabling direct, structured interactions between autonomous AI agents, supporting task coordination and collaboration ## Project Structure All the slides are located in the `slide/` directory. The code is located in the `examples/` directory. More infos about the examples is to be found in the [`MCP example README`](examples/mcp_example/README.md) and the [`A2A example README`](examples/a2a_example/README.md). ## AI Models The examples use OpenAI models by default (by importing the `langchain_openai` module for LangChain). You're welcome to switch to any model provider of your choice. ## Setup Instructions ### 1. Install Poetry Poetry is the dependency manager used in this project. Follow the [official installation guide](https://python-poetry.org/docs/#installation) to set it up on your system. ### 2. Install Project Dependencies ```bash poetry install ``` ### 3. Setup API keys Copy the file `.env.example` as `.env` and put your own keys. ### 4. Launch the applications #### 4.1 Streamlit MCP application ```bash poetry run streamlit run examples/mcp_example/app.py --server.address 0.0.0.0 ``` #### 4.2 A2A Application - Start the agent with the following command: ```bash poetry run python examples/a2a_example/agent/main.py --host 0.0.0.0 --port 8080 ``` - Then start the CLI client: ```bash poetry run python examples/a2a_example/client/main.py --agent http://localhost:8080 ```