Treasure Data MCP Server

No installation needed! Configure your MCP tool to run@treasuredata/mcp-serverdirectly vianpx:

What is npx?npx is a package runner that comes with npm 5.2+. It downloads and runs packages without installing them globally, ensuring you always use the latest version.

If you prefer a traditional installation:

Add to your MCP client configuration (e.g., Claude Desktop):

claude mcp add td -e TD_API_KEY=$TD_API_KEY -- npx @treasuredata/mcp-server

- Adds the server with the name "td"
- Sets the TD_API_KEY environment variable to your API key value
- Configures Claude Code to usenpx @treasuredata/mcp-server(always uses latest version)

You can also specify additional environment variables:

claude mcp add td \ -e TD_API_KEY=$TD_API_KEY \ -e TD_SITE=us01 \ -e TD_DATABASE=sample_datasets \ -- npx @treasuredata/mcp-server

Once configured, Claude Code will automatically have access to all the tools described below for querying and analyzing your Treasure Data.

List all databases in your Treasure Data account.

{ "name": "list_databases", "arguments": {} }

- database(string, optional): Database name. If omitted, uses the current database context (TD_DATABASE or last used database)

{ "name": "list_tables", "arguments": { "database": "sample_datasets" } }

{ "name": "list_tables", "arguments": {} }

Get schema information for a specific table.

- database(string, optional): Database name. If omitted, uses the current database context (TD_DATABASE or last used database)
- table(string, required): Table name

{ "name": "describe_table", "arguments": { "database": "sample_datasets", "table": "www_access" } }

{ "name": "describe_table", "arguments": { "table": "www_access" } }

Execute read-only SQL queries (SELECT, SHOW, DESCRIBE).

- sql(string, required): SQL query to execute
- limit(number, optional): Max rows (default: 40, max: 10000)

Performance Tip:For tables with atimecolumn, usetd_interval()ortd_time_range()to limit the time range:

- td_interval(time, '-30d/now')- Last 30 days
- td_interval(time, '-7d/now')- Last 7 days
- td_interval(time, '-1d')- Yesterday only
- td_interval(time, '-1h/now')- Last hour
- td_time_range(time, '2024-01-01', '2024-01-31')- Specific date range

{ "name": "query", "arguments": { "sql": "SELECT method, COUNT() as count FROM www_access GROUP BY method", "limit": 10 } }

{ "name": "query", "arguments": { "sql": "SELECT method, COUNT() as count FROM www_access WHERE td_interval(time, '-7d/now') GROUP BY method", "limit": 10 } }

Execute write operations (UPDATE, INSERT, DELETE, etc.) - requiresTD_ENABLE_UPDATES=true.

- sql(string, required): SQL statement to execute

{ "name": "execute", "arguments": { "sql": "INSERT INTO events (timestamp, event_type) VALUES (NOW(), 'test')" } }

Switch the current database context for subsequent queries.

- database(string, required): Database to switch to

{ "name": "use_database", "arguments": { "database": "production_logs" } }

After switching, all queries will use the new database by default unless explicitly specified.

Get the current database context being used for queries.

{ "name": "current_database", "arguments": {} }

{ "currentDatabase": "sample_datasets", "description": "The current database context used for queries" }

CDP Tools (Customer Data Platform) - EXPERIMENTAL

Note:CDP tools are currently experimental and may not cover all use cases. Additional functionality will be added based on user feedback.

The following tools are available for interacting with Treasure Data's Customer Data Platform (CDP):

List all parent segments in your CDP account.

{ "name": "list_parent_segments", "arguments": {} }

Get details of a specific parent segment.

- parent_segment_id(integer, required): The ID of the parent segment

{ "name": "get_parent_segment", "arguments": { "parent_segment_id": 12345 } }

List all segments under a specific parent segment.

- parent_segment_id(integer, required): The ID of the parent segment

{ "name": "list_segments", "arguments": { "parent_segment_id": 12345 } }

List all activations (syndications) for a specific segment.

- parent_segment_id(integer, required): The ID of the parent segment
- segment_id(integer, required): The ID of the segment

{ "name": "list_activations", "arguments": { "parent_segment_id": 12345, "segment_id": 67890 } }

Get detailed information about a specific segment, including its rules and metadata.

- parent_segment_id(integer, required): The parent segment ID
- segment_id(integer, required): The segment ID

{ "name": "get_segment", "arguments": { "parent_segment_id": 287197, "segment_id": 1536120 } }

Get the SQL statement for a parent segment.

- parent_segment_id(integer, required): The parent segment ID

{ "name": "parent_segment_sql", "arguments": { "parent_segment_id": 287197 } }

select a. from "cdp_audience_287197"."customers" a

{ "name": "segment_sql", "arguments": { "parent_segment_id": 287197, "segment_id": 1536120 } }

select a. from "cdp_audience_287197"."customers" a where ( (position('Male' in a."gender") > 0) )

Workflow Tools (Experimental) - Monitor and Control Digdag Workflows

Note: These workflow tools are experimental and provide detailed access to workflow sessions, attempts, and tasks. They are subject to change in future releases.

- limit(number, optional): Maximum results (default: 100)
- last_id(string, optional): Pagination cursor

{ "name": "list_projects", "arguments": { "limit": 50 } }

List workflows, optionally filtered by project name.

- project_name(string, optional): Project name to filter by
- limit(number, optional): Maximum results (default: 100)
- last_id(string, optional): Pagination cursor

// List all workflows { "name": "list_workflows", "arguments": { "limit": 50 } } // List workflows in a specific project { "name": "list_workflows", "arguments": { "project_name": "my_project", "limit": 50 } }