simple-streamable-mcp

aipachakutiqwan/simple-streamable-mcp

3.2

If you are the rightful owner of simple-streamable-mcp and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.

The Model Context Protocol (MCP) server is a framework designed to facilitate communication between clients and AI models, enabling rich-context interactions and efficient data processing.

simple-streamable-mcp

Welcome to the lightweight Streamable MCP client-server chatbot powered by Anthropic Claude as the LLM backend.

⚡ Introduction

This repository provides a simple implementation of a streamable MCP client-server chatbot that uses Anthropic Claude as the LLM backend. The chatbot is designed to answer questions about research papers on arXiv or extract specific information from them.

This implementation is a slight modification based on the publicly available MCP: Build Rich-Context AI Apps with Anthropic course.

:heavy_exclamation_mark: This repository code is not intended for direct use in production. Do not use it as-is. It serves as a reference implementation. You should review the logic, refactor the code to fit your requirements, and write appropriate unit tests before considering it for any production deployment.

:rocket: Setup

🌱 Create an environment and install dependencies

Check out the Contributing guide to learn how to set up the environment, install dependencies, and get started contributing to this repository.

🌱 Setting up env variables

You can use .env file for set the following enviromental variables.

export ANTHROPIC_API_KEY=SET_YOUR_API_KEY_HERE
export SERVER_CONFIG_FILE=config/server_config.yaml
export LOCAL_CONFIG_FILE=config/local_config.yaml
export LOG_CONFIG_FILE=config/log_config.yaml
export RUN_LOCALLY=False
export PORT=8001
export ANTHROPIC_MODEL=claude-3-7-sonnet-20250219
export MAX_TOKENS_MODEL=2024

🌱 Antrophic Claude API

  • Sign up here and set ANTHROPIC_API_KEY in your environment.

✅ How to Use

:pushpin: Locally

This version initializes the chatbot and creates an MCP instance locally using the local configuration file.

export RUN_LOCALLY=True
python src/client/mcp_chatbot.py

:pushpin: Client/Server

This version starts the MPC server and initializes an MCP client chatbot to communicate with it. Follow the next two steps to activate the Client/Server option.

Launch a terminal session and activate the MCP server.

export RUN_LOCALLY=False
python src/server/mcp_server.py

Open a new terminal and start the MCP client.

export RUN_LOCALLY=False
python src/client/mcp_chatbot.py

:pushpin: MCP Inspector

Enables activation of the MCP inspector interface for interaction with the MCP server.

Execute the following command, setting the STDIO parameters to establish a connection with the MCP server.

mcp dev src/server/mcp_server.py
  • Transport type: STDIO
  • Command: uv
  • Arguments: run --with mcp mcp run src/server/mcp_server.py

🤖 Chatbot

The chatbot will appear as shown below—this is an example of its interface in the terminal.

Hello! I'm here to help you search for academic papers on arXiv or extract information about specific papers.
(Write 'quit' to exit.)

Use: @papers to see available papers topics.
Use: @<topic> to search papers in that topic.
Use: /prompts to list available prompts.
Use: /prompt <prompt name> <arg1=value1> to execute a prompt.

Send a message: