Skip to main content

The AI-native analytics platform

Hex brings together the flexibility of code, the context of your business, and the tools for you to review, refine, and share what agents create.

How data teams at leading AI companies do their best work

Bring any agent to your data

Ask questions and build analyses from Hex, Claude, Codex, Cursor, or Slack.

Placeholder for a visual of multiple agents asking Hex questions

Ask and analyze from anywhere

Start a thread in Hex, mention @Hex in Slack, or call Hex from Claude, Codex, or Cursor through MCP and the CLI. Questions get answered wherever your team already works.

Placeholder for a visual of the agent's thinking steps

Every agent uses the same context on your data

Whichever agent asks, it reads the same definitions, endorsed projects, and business context — so the number in Slack matches the number in Cursor.

Placeholder for a visual of external agent conversations in Context Studio

See how the agent answered, and how it got there

Every answer shows its work: the projects it read, the cells it viewed, and the queries it ran. Conversations from external agents show up in Context Studio right next to the Hex Agent's.

Activate the context you already have

Placeholder for the Context Engine graphic

Sync the definitions you've already built, add the business details models miss, and let context improve with every answer.

  • Sync from anywhere

    Bring in dbt, Snowflake semantic views, warehouse metadata, Markdown files, and GitHub repos.

  • Start fast

    Begin from the signals already in your warehouse. A semantic layer helps, but you don't need one to get going.

  • Self-improving

    Hex flags context gaps and suggests fixes, so answers get sharper the more your team asks.

  • Portable

    Your context works across models and warehouses, with no lock-in. Switch either one and it comes with you.

Trust agent-generated apps

Placeholder for a shared Hex app with comments, share settings, and a SQL diff

Every Hex project is code that agents can write and people can review. Inspect the analysis, then publish it as an interactive app your whole team can explore.

Governed from the warehouse up

Every question and answer inherits your identity, permissions, and audit requirements, whichever agent started it.

Identity & access

SSO and SCIM, plus permissions inherited straight from your warehouse.

Observability

Every agent interaction is visible in Context Studio, wherever the question was asked.

Audit & compliance

Audit logs and SIEM export, with SOC 2, HIPAA, GDPR, and VPC support.

Cost controls

Pick which models agents use, set credit limits, and track usage across the workspace.

Works with your whole stack

Placeholder for a layer cake of integration logos by category

Hex builds on the tools you've already invested in, so every agent works from the same governed context.

  • AI agents and clients

    Claude, Codex, Cursor, and ChatGPT connect through MCP and the CLI, and reach across every layer below.

  • Analytics, dashboards, and notebooks

    Hex covers this layer in one platform: chat, notebooks, dashboards, and apps.

  • Transformation and semantic layer

    dbt, Snowflake semantic views, and Cube sync into Hex as agent context.

  • Warehouses and lakehouses

    Snowflake, BigQuery, Databricks, Redshift, Athena, and Postgres.

  • Ingestion and orchestration

    Fivetran, Airbyte, Airflow, and Dagster. Hex works on the data your pipelines deliver.

Give every agent the context it needs

See where Hex fits in your AI stack.

FAQ

You don't have to pick one. On their own, general-purpose assistants don't know your metric definitions, which tables to trust, or who's allowed to see what. Connect them to Hex through MCP or the CLI and they work from the same governed context as the Hex Agent. Every answer shows the queries behind it, and the work can be reviewed, refined, and published as an app your team can explore.

Warehouse AI features are built around the data in that warehouse. Hex works one layer up: it syncs context from dbt, semantic views, warehouse metadata, Markdown files, and GitHub repos, and that context works across models and warehouses. Every answer also gets a home — an analysis you can inspect and an app you can share. Hex inherits your warehouse permissions, so governance stays where you already manage it.

Admins choose which models agents use, set credit limits, and track usage. For accuracy, every agent works from the same context, and Context Studio shows every agent interaction — including questions asked from Slack, Claude, or Cursor. The Review Agent flags context gaps and suggests fixes, and you can test context changes with evals before they ship.

For many teams, yes. Hex brings chat, notebooks, dashboards, and apps into one platform, so analysis doesn't sprawl across a BI tool, a notebook, and a dozen screenshots in Slack. It sits on top of your warehouse and transformation layer rather than replacing them.