context serve

The serve command starts a Model Context Protocol (MCP) server that makes your compiled context available to AI assistants and other MCP clients.

Usage

context serve

Option	Description	Default
`--http`	Use HTTP transport (default is stdio)	Off
`--port <number>`	Port for HTTP server	`3000`
`--host <address>`	Host address to bind	`0.0.0.0`
`--context-dir <path>`	Path to context files	`./context`

context serve

This is the standard mode for use with Claude Desktop, Cursor, and other MCP-compatible tools.

context serve --http --port 8080

Starts an HTTP server at http://0.0.0.0:8080/mcp for browser-based or networked clients.

context serve --http --host 0.0.0.0 --port 3000

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "context": {
      "command": "context",
      "args": ["serve"],
      "cwd": "/path/to/your/project"
    }
  }
}

Add to .cursor/mcp.json in your project:

{
  "mcpServers": {
    "context": {
      "command": "context",
      "args": ["serve"]
    }
  }
}

The server exposes the following to MCP clients:

Resource	Description
`context://manifest`	Full compiled manifest
`context://model/{name}`	Single model details
`context://glossary`	All business term definitions
`context://tier/{name}`	Tier scorecard for a model
`context://data-product/template`	Blank data product YAML template
`context://data-product/{name}`	Export model as data product

Tool	Description
`context_search`	Full-text search across models, fields, terms
`context_explain`	Look up any entity by name
`context_validate`	Run lint rules and return diagnostics
`context_tier`	Get the current tier scorecard
`context_golden_queries`	Retrieve curated SQL templates
`context_guardrails`	Get required filters and business rules