Cube

Universal semantic layer for building data applications with consistent metrics.

MCPCLI

About Cube

Universal semantic layer for building data applications with consistent metrics. Explore how Cube integrates with the agentic data stack ecosystem and supports autonomous data operations.

Key Features

  • Code-first data modeling using YAML or JavaScript with version control integration
  • Intelligent caching via pre-aggregations in Cube Store for aggregate-aware query acceleration
  • Multiple API standards: REST, GraphQL, SQL, MDX, and DAX for universal tool interoperability
  • AI API providing a standard interface for LLM-powered text-to-semantic-layer queries
  • Comprehensive access control with Python/JavaScript-defined security policies
  • Semantic SQL architecture (Postgres-compatible with MEASURE function) for AI agents
  • Dataset-centric design using Cubes and Views as a queryable knowledge graph

Agent Integration

CLI — cubejs-cli

$ npx cubejs-cli <command> OR npm install -g cubejs-cli
CLI Documentation

External Links

Cube MCP Server Documentation

Official docs for Cube's MCP server integration covering OAuth, Chat API, and setup

REST API Reference

REST API docs for programmatic querying of the semantic layer with OpenAPI spec

AI API Practical Guide

Guide to the native AI API for building agents, chatbots, and copilots on top of Cube

JavaScript SDK

Client SDK for building applications that query the Cube semantic layer

Community Cube MCP Server

Community-built MCP server for interacting with Cube semantic layers

← Back to Semantic Layer