Overview
What is Nf-core tools MCP (Unofficial)?
This is an experimental MCP (Model Context Protocol) server that provides programmatic access to key nf-core tools functions. It allows users to create pipelines, list and create modules, and install existing modules—all through MCP-compatible clients like Claude Desktop or Cursor. It is designed for nf-core developers and bioinformaticians who want to integrate nf-core workflows into AI-assisted environments.
How to use Nf-core tools MCP (Unofficial)?
Installation requires Claude Desktop or Cursor, Uv, and nf-core tools. Create a virtual environment with uv venv, activate it, then add packages with uv add "mcp[cli]" nf-core. Finally, install one of the two server files using fastmcp install nf-core.test-server.cli.py (CLI-based) or fastmcp install nf-core.test-server.python.py (Python-based). To add the server to Cursor, configure a JSON entry pointing to the repo directory with uv run server.py.
Key features of Nf-core tools MCP (Unofficial)
- Create a pipeline using the nf-core template
- List available nf-core modules
- Create a module using the nf-core modules template
- Install existing nf-core modules into a pipeline or modules repo
- Two implementations: CLI-based and Python-based
- Integrates with Claude Desktop and Cursor via MCP
Use cases of Nf-core tools MCP (Unofficial)
- Bootstrapping a new nf-core pipeline without leaving an MCP client
- Browsing the nf-core modules registry and installing modules into an existing pipeline
- Automating module creation during pipeline development workflows
- Experimenting with MCP capabilities for nf-core tooling in AI-assisted environments
FAQ from Nf-core tools MCP (Unofficial)
What does Nf-core tools MCP (Unofficial) do?
It provides MCP-compatible tools that wrap nf-core command-line functions, enabling users to create pipelines, list and create modules, and install modules directly from MCP clients.
What are the prerequisites for using this server?
You need Claude Desktop or Cursor, Uv, and nf-core tools installed on your system.
How do I install the server?
Create a virtual environment with uv venv, activate it, add the required packages (uv add "mcp[cli]" nf-core), then run fastmcp install with either the CLI or Python server file.
What are the two server implementations?
The CLI-based server (nf-core.test-server.cli.py) calls nf-core commands via subprocess, while the Python-based server (nf-core.test-server.python.py) uses the nf-core Python package directly. Both provide identical tools.
Is this an official nf-core project?
No, this is an unofficial experiment by the repository author to explore MCP capabilities for nf-core tools functions.
