Decoupler MCP Server

What is decoupler-MCP?

decoupler-MCP is a natural language interface for scRNA-Seq analysis that works with decoupler through the Model Context Protocol (MCP). It is designed for anyone who wants to perform single-cell RNA sequencing analysis using natural language, as well as agent developers who need to call decoupler’s functions programmatically.

How to use decoupler-MCP?

Install the package via pip (pip install decoupler-mcp) and test it by running decoupler-mcp run. For local use, configure your MCP client with the command path and arguments; for remote use, run the server with --transport shttp --port 8000 and point your client to the corresponding URL. The server is supported in AI clients (e.g., Cherry Studio), plugins (e.g., Cline), and agent frameworks (e.g., Agno).

Key features of decoupler-MCP

• IO module for reading and writing scRNA-Seq data
• Pathway activity and transcription factor inference
• Tool module including clustering and differential expression
• Plotting module with violin, UMAP, and t-SNE visuals
• Runs locally or remotely via SSE HTTP transport
• Plugs into multiple MCP-compatible AI clients and frameworks

Use cases of decoupler-MCP

• Perform whole scRNA-Seq analysis pipelines using natural language commands
• Integrate decoupler’s biological activity inference into AI agent workflows
• Automate clustering and differential expression analysis without coding
• Visualize single-cell data with UMAP, t-SNE, or violin plots interactively