Hex homepage

Hex

by Hex Technologies

Collaborative SQL and Python notebooks with AI agents for trusted analytics

Product with AI agentsCopilot

Last reviewed 2026-06-18

Visit websiteDocsAlternatives

Hex is a collaborative data analytics platform built around notebooks that mix SQL, Python, and no-code cells. Data teams query warehouses, build interactive charts and dashboards, and ship results as shareable data apps. It sits on top of an existing warehouse rather than storing data, and layers governance, version history, and a semantic model so business users can self-serve trusted answers. Over the last two years Hex added AI agents on top of the notebook. The headline Notebook Agent discovers data sources, plans an analysis, writes and runs SQL, Python, chart, and markdown cells, and summarizes findings, while requiring the analyst to confirm or undo every change at the cell level. Companion features include Threads (conversational self-serve in Hex or Slack), a modeling agent, and an MCP server that lets external assistants query a governed workspace.

What it can do

  • Draft and run notebook analyses (Notebook Agent)

    Copilot

    Discovers tables, builds a plan, and writes and executes SQL, Python, chart, pivot, and markdown cells. Every change is staged as a pending change the analyst must explicitly confirm or undo, with inline diffs.

    source

  • Answer data questions in natural language (Threads)

    Assistant

    Conversational self-serve in Hex or Slack that prioritizes endorsed, semantically modeled data and returns answers with lineage and context.

    source

  • Generate semantic models and data apps

    Copilot

    Generates and edits semantic models and builds interactive data apps from a prompt, for technical users who can audit the generated code.

    source

  • Expose the workspace to external AI assistants (MCP server)

    Assistant

    An MCP server lets external clients like Claude, Cursor, and ChatGPT search projects and create or continue Threads against a governed workspace.

    source

Strengths

  • +Unified SQL + Python + no-code notebook with strong collaboration, version history, and a clean path from analysis to shareable data app
  • +Warehouse-native and broadly integrated (Snowflake, BigQuery, Databricks, Redshift, dbt)
  • +Governance-friendly, conservative AI: cell-level confirm/undo with diffs keeps a human in the loop

Limitations

  • −Per-editor pricing adds up, and the most useful AI and governance features are gated to Team and Enterprise tiers
  • −Several AI features (MCP server, generative apps) are still in beta
  • −The Notebook Agent is aimed at technical users who can audit code; deepest value assumes SQL or Python literacy

Overview

Hex is a collaborative data analytics platform built around notebooks that mix SQL, Python, and no-code cells, with shareable data apps and a governance layer. It is an established platform (founded 2019) that has layered multiple AI agents on top of the notebook.

What it does

The headline Notebook Agent discovers data sources, plans an analysis, and writes and runs cells, staging every change for cell-level confirm or undo. Threads offers conversational self-serve Q&A in Hex or Slack, prioritizing semantically modeled data. A modeling agent and generative apps round out the AI surface, and an MCP server exposes the governed workspace to external assistants.

Integrations & setup

Warehouse-native across Snowflake, BigQuery, Databricks, Redshift, Postgres, and more, with dbt and Cube metrics, GitHub and GitLab version control, Slack, and orchestration tools. It offers a REST API and an MCP server (beta) compatible with Claude, Cursor, and ChatGPT.

Pricing

Freemium: a free Community tier, Professional at about $36/editor/mo, Team at about $75/editor/mo, and custom Enterprise. The deepest AI and governance features live in Team and Enterprise.

Best for / not for

Best for mid-market and enterprise data teams that want a governed, collaborative notebook with a careful AI layer. Less suited to non-technical users who want a pure plain-English tool, where the agent's audit-the-code posture adds friction.

Alternatives

Julius AI targets the non-technical, plain-English end of data analysis.

What people are saying

We aggregate real LinkedIn discussion into sentiment for the agents people search most. Hex isn't tracked yet, want it added? Request tracking.

FAQ

Is the Hex Notebook Agent autonomous?+

No. It drafts and can run cells, but every change is staged as a pending edit the analyst must explicitly confirm or undo, with inline diffs. Hex's docs state it is intended for technical users who can audit the generated SQL and code.

Does Hex store my data?+

No. Hex sits on top of your existing warehouse, querying in-database or caching to a dataframe, rather than storing your data.

Sources

Last reviewed 2026-06-18

Alternatives & related

Julius AI
AI data analyst: connect data, ask in plain English, get analysis and charts