Loading tool…
Semantic layer for analytics: define a metric once, then query it from dashboards, embedded apps and AI agents
Cube sits between a warehouse and everything that asks it questions. You model metrics and dimensions once in code, and Cube serves those same definitions to dashboards, workbooks, embedded customer-facing analytics, an MCP server that Claude, ChatGPT or Cursor can call, and a Slack agent, with pre-aggregation caching underneath for speed. It connects to Snowflake, BigQuery and Databricks. Cube Dev, Inc. runs it from San Francisco and also maintains the open-source Cube Core project. Cube Cloud has a free tier; paid plans start at $40 per developer per month, and Enterprise adds single-tenant infrastructure, BYOC and SAML SSO.
Free tier covers data source connections, modelling, workbooks, dashboards and agents. Starter is $40 per developer per month and adds production deployment compute, Cube Store caching and premium models; Premium is $80 per developer per month and adds embedded dashboards, embedded chat, viewer roles and a 99.95% uptime SLA. Enterprise is quoted.
Use tool ↗Cube's value is that one metric definition serves a dashboard, a customer-facing embed and an AI agent identically, which is exactly where bolt-on chat BI falls apart. The catch is the shape of the bill: it is per developer, and the features most teams actually want from a semantic layer - embedded analytics and embedded chat - sit on the $80 tier rather than the $40 one.
Watch out: Embedded analytics and the embedded chat widget are Premium ($80/developer) features, not Starter ($40) ones - check which tier your actual use case needs before budgeting.
No reviews yet — be the first to review Cube.
Reviews are tied to your EffectHub account — one review per tool, so the rating for Cube reflects real users.
No questions yet — ask the first one about Cube.
Questions about Cube are tied to your EffectHub account, so answers can reach you and the section stays free of spam.