Lightdash MCP Server

Instructions

Implementation Reference

• src/mcp.ts:439-469 (handler)

  Handler function that parses input arguments using GetAnalyticsRequestSchema, calls the Lightdash API endpoint '/api/v1/projects/{projectUuid}/dataCatalog/{table}/analytics' with projectUuid and table parameters, handles errors, and returns the results as JSON text.

  case 'lightdash_get_analytics': {
    const args = GetAnalyticsRequestSchema.parse(
      request.params.arguments
    );
    const { data, error } = await lightdashClient.GET(
      '/api/v1/projects/{projectUuid}/dataCatalog/{table}/analytics',
      {
        params: {
          path: {
            projectUuid: args.projectUuid,
            table: args.table,
          },
        },
      }
    );
    if (error) {
      throw new Error(
        `Lightdash API error: ${error.error.name}, ${
          error.error.message ?? 'no message'
        }`
      );
    }
    return {
      content: [
        {
          type: 'text',
          text: JSON.stringify(data.results, null, 2),
        },
      ],
    };
  }

• src/schemas.ts:96-104 (schema)

  Zod schema for input validation: requires projectUuid (UUID string with description) and table (string).

  export const GetAnalyticsRequestSchema = z.object({
    projectUuid: z
      .string()
      .uuid()
      .describe(
        'The UUID of the project. You can obtain it from the project list.'
      ),
    table: z.string(),
  });

• src/mcp.ts:104-108 (registration)

  Registration of the 'lightdash_get_analytics' tool in the MCP server's tool list, specifying name, description, and input schema derived from GetAnalyticsRequestSchema.

  {
    name: 'lightdash_get_analytics',
    description: 'Get analytics for a specific table in the data catalog',
    inputSchema: zodToJsonSchema(GetAnalyticsRequestSchema),
  },