mcp-multi -anthropic-chatbot

This MCP server is a multi-server chatbot designed to help you discover and analyze arXiv research papers using natural language queries. It connects to multiple MCP servers simultaneously, exposing a range of tools such as paper search, content extraction, and topic-based browsing. You can issue natural language prompts to search for papers, fetch and summarize contents, and leverage pre-built prompts to structure complex analyses. The system integrates a research-oriented server for arXiv interactions, a fetch server for web content retrieval, and a filesystem server for managing files and resources. Use the provided commands to interact with these servers and to access organized paper collections by topic.

Prerequisites:

• Python 3.8+ installed on your system
• A virtual environment tool (optional but recommended)
• Access to an Anthropic API key

1. Create and activate a virtual environment (optional but recommended):

python3 -m venv .venv
source .venv/bin/activate

2. Install dependencies listed in requirements.txt:

pip install -r requirements.txt

3. Set up your API key for Anthropic in an environment variable or a .env file:

export ANTHROPIC_API_KEY=your_api_key_here

4. Run the MCP chatbot server using the uv runner (as shown in the README):

uvx run mcp_chatbot.py

5. Optional: if using a script-managed environment, ensure your environment variables are loaded before starting the server.

Notes:

• The server relies on multiple components described in the README (Research Server, Fetch Server, Filesystem Server). Ensure those resources are accessible when testing.
• If you encounter permission or network issues, verify your API keys and network access to arXiv and any external services.

Tips and considerations:

• The Anthropic API key is required for the anthropic-based prompts and analyses.
• The system uses a multi-server setup; ensure all referenced servers are available and properly configured in server_config.json if you adapt or extend the configuration.
• Use the provided prompts and commands to search papers, browse topics, and generate analyses. For example, you can request a search prompt with topic=neural networks and a specific number of papers, then fetch and summarize the results.
• If you modify the codebase, keep an eye on the mcp_chatbot.py entrypoint to ensure compatibility with the uvx runner.