Hex
Free tierThe AI analytics platform where trust meets insight — for your whole team
Free tier available·All audiences·Powered by OpenAI·API available
Key strengths
Agentic AI that writes SQL & Python and builds charts from natural language promptsSemantic model layer that grounds AI answers in trusted, business-defined metricsDual interface: collaborative data notebooks plus polished, shareable data appsConversational self-serve analytics so non-technical stakeholders can explore data without writing codeFlexible context approach that earns trust without slowing down analysts
Free tier + paid plans
San Francisco, USA
Founded 2019
No ratings yet
Developer & Analyst Documentation
Hex is built for data practitioners who want maximum flexibility alongside powerful AI assistance.
- Notebooks: Hex notebooks support both SQL cells and Python cells side-by-side. Variables are shared across cells, making it easy to pass query results into pandas, Plotly, or any other Python library.
- Semantic models: Define endorsed semantic models with standardized metrics, dimensions, and filters. The Magic AI agent automatically selects the most relevant endorsed model when answering questions, ensuring AI-generated analysis is grounded in governed business logic rather than raw tables.
- Agentic workflows: The Magic agent reasons over your schema, selects data sources, writes and executes code, and iterates — similar to an autonomous coding agent but purpose-built for analytics. Thought traces are visible for transparency.
- App builder: Convert any notebook into an interactive data app with dropdowns, filters, and charts. Apps can be published and shared without requiring viewers to have a Hex account.
- Integrations & scheduling: Connect to warehouses (Snowflake, BigQuery, Databricks, Redshift), schedule notebook runs, manage packages and files, and inspect full run history via the workspace UI or API.
- Conversational self-serve: End users can interact with published apps through a chat interface powered by the same semantic models, enabling ad-hoc exploration without analyst intervention.