This MCP server provides a FastMCP-compatible interface for searching and extracting research papers from arXiv. It is designed for local development with uv (the Python dependency manager) and deployment on hosting platforms like Render. Once running, you can interact with the server to submit search queries against arXiv, retrieve metadata, and extract document details for selected papers. The server leverages arXiv data sources to return structured results that you can feed into downstream workflows or dashboards.

Prerequisites:

• Python 3.11.x (the project references Python 3.11.11 on Render)
• uv (Python dependency manager)
• git

Installation steps:

1. Clone the repository: git clone https://github.com/your-org/mcp-research cd mcp-research

2. Install dependencies using uv (from the project root): uv install -r requirements.txt

   or if you prefer explicit installation via uv:

   uv pip install -r requirements.txt

3. Run the development server locally: uv run research_server.py

4. (Optional) Regenerate requirements.txt after changes to pyproject.toml: uv pip compile pyproject.toml --no-emit-find-links > requirements.txt

Note: For Render deployment, the runtime configuration handles dependencies from requirements.txt and uses Python 3.11.11 as specified in render.yaml.

Tips and notes:

• The server relies on uv for dependency management; ensure requirements.txt is kept up-to-date if you modify pyproject.toml.
• Render deployments expect the port to be provided by the environment; do not hard-code a port in the server startup.
• If you encounter import errors, run uv install or uv pip install -r requirements.txt to refresh dependencies.
• Since this server targets arXiv data, consider handling API rate limits and implementing caching in external clients to optimize repeated queries.
• The npm_package field is null because this is a Python-based server, not Node.js.