Skip to content
Best Data Catalogs mark: an unfinished lineage pathBest Data Catalogs

Platform review

Semantic layer

Cube review: a semantic layer for AI agents

Short answer

Cube is an agentic analytics platform built on a semantic layer, with an MCP server included in its free plan and paid plans from $40 per developer per month. It scores 5.86 and ranks 7 of 9. It is strong on metric definitions and pricing transparency and does not cover catalog lineage.

By The Lineage Desk · Best Data Catalogs · Published

Desk research from public vendor pages. Editorial assessment.

5.86 / 10

Rank 7 of 9 on default weights. Change the weights

What does Cube do for AI agents?

Cube is an agentic analytics platform built on a semantic layer: people and AI agents query the same governed model through BI, Analytics Chat, MCP and APIs.

Agent interface
MCP server, APIs, Analytics Chat API
Lineage
Not described (a semantic layer, not a catalog)
Deployment
Cube Core is open source (cube.dev shows 21,018 GitHub stars); Cube Cloud is the managed platform.
Best for
Teams whose main gap is consistent metrics for agents and BI, and who want published pricing and an open-source core

Which hops from warehouse to agent does Cube document?

Cube context path: documents 4 of 5 stages.1Warehouse and lakehouse2Transformation (dbt, ETL)3BI and semantic logic4Governance and access rules5Delivery to agents (MCP or API)
  • Solid dotDocumented on the vendor's pages
  • Dashed hollow dotNot documented on the pages we read

  1. 1Warehouse and lakehouse: Documented

    Runs on cloud data platforms; Cube + Snowflake CoCo.

    Source: Cube homepage

  2. 2Transformation (dbt, ETL): Not documented

    Transformation tools are not described on the pages we read.

  3. 3BI and semantic logic: Documented

    BI for people and agents from one semantic model.

    Source: Cube homepage

  4. 4Governance and access rules: Documented

    Governed data served through MCP and APIs.

    Source: Cube homepage

  5. 5Delivery to agents (MCP or API): Documented

    MCP server and APIs.

    Source: Cube homepage

A filled node means the vendor's own pages describe that hop. It does not measure how well it works.

What are Cube's key features?

  • A semantic layer that serves BI for people and AI agents from one governed model.

    Source: Cube homepage

  • An MCP server for AI agents, included in the free plan.

    Source: Cube pricing

  • Analytics Chat for natural-language questions against the semantic model; the Analytics Chat API is agent-to-agent capable via MCP.

    Source: Cube homepage

  • Cube + Snowflake CoCo, announced 17 September 2026.

    Source: Cube homepage

What does Cube claim?

What does Cube cost?

Free ("free forever", includes an MCP server for AI agents, analytics chat and a semantic model agent); Starter $40 per developer per month; Premium $80 per developer per month; Enterprise custom.

Entry price: Free; paid plans from $40 per developer per month

Source: Cube pricing

How does Cube score on each criterion?

CriterionWeightScoreWhy
Agent access to context15%8 / 10Governed data reaches agents through MCP and APIs; the free plan includes an MCP server.

Source: Cube pricing

Cross-stack lineageTied lowest13%2 / 10A semantic layer and BI platform, not a catalog; cross-stack lineage is not described.

Source: Cube homepage

Automated context build13%5 / 10A semantic model agent is listed; automated discovery and documentation across the stack are not described.

Source: Cube pricing

Governance agents can useLowest in set13%5 / 10Describes governed data for agents; certification and sensitivity rules are not detailed.

Source: Cube homepage

Business definitions and metricsTied highest12%9 / 10Built on a semantic layer that people and agents query alike.

Source: Cube homepage

Answer quality checksTied lowest8%2 / 10Checks on agent answers are not described on the pages read.

Source: Cube homepage

Integration breadthTied lowest10%5 / 10Works on cloud data platforms and names agent clients; no connector count was read.

Source: Cube homepage

Pricing transparencyTied highest8%9 / 10Published: Free, Starter $40 and Premium $80 per developer per month, Enterprise custom.

Source: Cube pricing

Deployment options8%8 / 10Open-source core (21,018 GitHub stars shown on cube.dev) alongside managed Cube Cloud.

Source: Cube homepage

Where does it score well, and where does it fall short?

Scores well

Highest in set on: Semantics and Pricing.

  • Published, low entry pricing, with an MCP server on the free plan.
  • An open-source core you can run yourself.
  • Agents and dashboards answer from the same metric definitions.

Falls short

Lowest in set on: Lineage, Governance, Answer checks and Integrations.

  • No catalog and no cross-stack lineage.
  • Answer testing and automated discovery are not described.
  • Governance rules for agents are not detailed on the pages we read.

Key facts

Company
cube.dev shows 21,018 stars on GitHub and 12,830 members on Slack.

Source: Cube homepage

Customers named
Named on Cube's site: Brex, Alcon, Drata and Webflow.

Source: Cube homepage

How does Cube compare head to head?

Cube alternativesBuild a shortlist with Cube

Frequently asked questions

What does Cube cost?

Free ("free forever", includes an MCP server for AI agents, analytics chat and a semantic model agent); Starter $40 per developer per month; Premium $80 per developer per month; Enterprise custom.

Does Cube have an MCP server?

Yes. Cube's free plan includes an MCP server for AI agents.

Can Cube be self-hosted?

Cube Core is open source (cube.dev shows 21,018 GitHub stars); Cube Cloud is the managed platform.