MCP Server Setup

DataSpoc exposes MCP (Model Context Protocol) servers for both Lens and Pipe. This lets AI assistants discover tables, run queries, manage pipelines, and more --- directly from the conversation.

Lens MCP Server

Install

pip install dataspoc-lens[mcp]

Run

dataspoc-lens mcp

Claude Desktop Configuration

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "dataspoc-lens": {
      "command": "dataspoc-lens",
      "args": ["mcp"],
      "env": {
        "DATASPOC_BUCKET": "s3://my-data"
      }
    }
  }
}

Available Tools

Tool	Description
list_tables	List all tables in the data lake
describe_table	Get schema, row count, and partitions for a table
query	Execute a SQL query and return results
ask	Ask a natural language question about the data
cache_status	Show cache freshness and size
cache_refresh	Refresh all cached data from the bucket
cache_refresh_stale	Refresh only tables with stale cache

Available Resources

Resource	Description
lens://tables	Full table catalog with schemas and metadata

Pipe MCP Server

Install

pip install dataspoc-pipe[mcp]

Run

dataspoc-pipe mcp

Claude Desktop Configuration

{
  "mcpServers": {
    "dataspoc-pipe": {
      "command": "dataspoc-pipe",
      "args": ["mcp"],
      "env": {
        "DATASPOC_BUCKET": "s3://my-data"
      }
    }
  }
}

Available Tools

Tool	Description
list_pipelines	List all configured pipelines
pipeline_config	View configuration for a pipeline
run_pipeline	Execute a pipeline
pipeline_status	Check run status and last sync time
pipeline_logs	View execution logs
show_manifest	Read the bucket manifest
validate_pipeline	Validate pipeline configuration

Available Resources

Resource	Description
pipe://pipelines	Full pipeline catalog with configurations

Using Both Servers Together

You can run both MCP servers simultaneously. Add both to your Claude Desktop config:

{
  "mcpServers": {
    "dataspoc-lens": {
      "command": "dataspoc-lens",
      "args": ["mcp"],
      "env": {
        "DATASPOC_BUCKET": "s3://my-data"
      }
    },
    "dataspoc-pipe": {
      "command": "dataspoc-pipe",
      "args": ["mcp"],
      "env": {
        "DATASPOC_BUCKET": "s3://my-data"
      }
    }
  }
}

This gives the agent full control: ingest new data with Pipe, then query it with Lens --- all from a single conversation.

Cursor and Windsurf

Both Cursor and Windsurf support MCP servers. Add the same configuration to their respective settings files. Refer to the Cursor or Windsurf documentation for the exact config file location.