Cursor

Setup

Cursor supports MCP servers through its configuration. Add BioContext7 to your Cursor MCP settings.

Local stdio

Create or edit .cursor/mcp.json in your project root:

{
  "mcpServers": {
    "biocontext7": {
      "command": "npx",
      "args": ["-y", "@biocontext7/mcp"]
    }
  }
}

Remote HTTP

When api.biocontext7.com is deployed:

{
  "mcpServers": {
    "biocontext7": {
      "url": "https://api.biocontext7.com/mcp"
    }
  }
}

Usage

Once configured, BioContext7 tools are available in Cursor's AI chat:

• Ask about bioinformatics tools and Cursor will use resolve-library-id
• Request documentation for a specific tool with get-library-docs
• Get ranked pipeline recommendations with recommend-tools

Example Prompts

"Find tools for single-cell analysis"
"Get documentation for STAR aligner"
"Suggest a pipeline for metagenomics analysis"
"What tools do differential expression analysis?"

Cursor Rules

The BioContext7 Cursor plugin includes bio-specific rules that enhance AI assistance for pipeline development.

.cursorrules

Copy the .cursorrules file to your project root for project-level bioinformatics context:

cp plugins/cursor-biocontext7/.cursorrules /path/to/your/project/

This gives Cursor context about available MCP tools, standard analysis chains, container conventions, and EDAM data formats.

.cursor/rules/*.mdc

Copy the conditional rules for file-type-specific guidance:

cp -r plugins/cursor-biocontext7/.cursor/rules/ /path/to/your/project/.cursor/rules/

Available rules:

• biocontext7.mdc — Activates for *.nf, *.smk, *.wdl, *.cwl, *.py files
• pipeline-patterns.mdc — Standard analysis chains and resource defaults for workflow files

Verify

1. Open Cursor with the MCP config in place
2. Check that the BioContext7 MCP server appears in the MCP panel
3. Test with: "find tools for single-cell analysis" — the resolve-library-id tool should fire

HTTP Transport

For persistent connections, use HTTP transport:

1. Start the BioContext7 server:

npx @biocontext7/mcp --transport http --port 3000

2. Configure Cursor to connect:

{
  "mcpServers": {
    "biocontext7": {
      "url": "http://localhost:3000/mcp"
    }
  }
}