Hex

The Notebook Agent accepts natural language queries and generates executable SQL code by searching endorsed semantic models and table schemas in connected data warehouses. The agent serializes notebook context (available tables, previous queries, semantic definitions) and uses an LLM to synthesize SQL that references specific tables and metrics by name, then executes the generated code server-side on Hex infrastructure with configurable compute profiles (Small to 4XL CPU/GPU options).

The Notebook Agent generates executable Python code from natural language requests by analyzing the current notebook state (previous cell outputs, imported libraries, variable definitions) and synthesizing code that integrates with existing analysis. Generated code executes server-side on Hex compute infrastructure, with access to standard Python libraries and the ability to reference upstream cell outputs as DataFrames or other objects.

Published apps (Team+ feature) support visual data exploration where users can drill down into underlying data by clicking on chart elements or table rows. The system automatically generates drill-down queries based on the selected data point, enabling users to explore data hierarchies without manual query writing. Drill-down is only available in published apps, not in edit mode.

Hex offers six compute tiers (Small: 2GB RAM/0.25 CPU through 4XL: 96GB RAM/24 CPU) plus optional GPU acceleration. Free tier limited to Small compute; Medium compute (8GB RAM/1 CPU) included on all paid plans; Large+ tiers incur per-minute charges ($0.32-$2.58/hr for CPU, $2.93-$4.06/hr for GPU). Users select compute profile per notebook, and costs are billed per-minute of execution time beyond included allowances.

Team+ tier enables exporting notebooks as Git projects and importing packages (shared components, templates) from other notebooks. This allows teams to version control notebooks in Git, share reusable components across projects, and maintain a library of analysis templates. Export format and Git integration details not fully documented.

Enterprise plan includes OIDC single sign-on (SSO) for centralized authentication, OAuth database connections for warehouse access, audit logs for compliance tracking, and HIPAA compliance certification. These features enable organizations to enforce authentication policies, track user actions, and meet regulatory requirements without managing credentials in Hex.

Enterprise plan enables embedding Hex apps in external websites (embedded analytics) and deploying custom Docker images with pre-installed packages or custom runtime environments. Single-tenant deployment option available for organizations requiring isolated infrastructure.

Enterprise plan option for deploying Hex in a single-tenant environment with HIPAA compliance, custom branding (white-label), and dedicated support. Enables embedding Hex analytics in customer-facing applications without Hex branding. Requires custom contract and pricing.

Hex notebooks use a reactive execution model where cells (SQL, Python, or no-code) automatically re-execute when their dependencies change, rather than executing sequentially top-to-bottom like Jupyter. The system tracks data flow between cells and intelligently re-runs only affected downstream cells when an upstream cell is modified, enabling fast iteration without manual re-execution. Execution happens server-side on configurable compute profiles (Small: 2GB RAM / 0.25 CPU through 4XL: 96GB RAM / 24 CPU, with optional GPU).

When a SQL or Python cell produces tabular results, Hex automatically generates visualizations (bar, line, pie, scatter charts inferred from data types and cardinality) without requiring explicit chart configuration. Users can customize chart type, axes, colors, and aggregations through a visual editor. The system also supports pivot tables, summary statistics, and drill-down exploration in published apps, with all visualizations rendered client-side for fast interaction.

Hex provides a no-code app builder that allows users to drag components (charts, tables, input controls, text) onto a canvas to create interactive dashboards without writing HTML/CSS/JavaScript. Apps support parameterization through input controls (dropdowns, date pickers, text inputs) that filter underlying queries, and can be published with permission controls (view-only or edit-restricted). Published apps render in a browser and support drill-down data exploration (Team+ feature) and email/Slack alerts (Team+ feature).

Team+ tier enables scheduling notebooks to run on a recurring basis (daily, weekly, monthly, or custom cron schedules) and send results via email or Slack. When a notebook runs on schedule, all cells execute in dependency order, and results can be formatted as email summaries or Slack messages. Scheduled runs use the same compute infrastructure as manual execution, with costs billed per-minute for compute usage beyond the included Medium tier.

Hex notebooks support multiplayer real-time editing where multiple users can edit cells simultaneously (conflict resolution mechanism not specified). Comments can be attached to individual cells or published app outputs, creating separate discussion threads. Version history is retained for 7 days (free), 30 days (Professional), or unlimited (Team+), allowing users to revert to previous notebook states. Shared components and collections (Team+ feature) enable code reuse across notebooks.

Hex integrates with dbt semantic models, allowing notebooks to reference endorsed metrics and standardized table definitions. The Semantic Model Agent (Team+ feature) can answer questions about metrics and generate queries that use pre-calculated metrics rather than raw tables. SQL cells can reference semantic model metrics directly, and the agent is aware of metric definitions when generating code, ensuring queries use business logic rather than raw calculations.

Hex supports connections to major cloud data warehouses (Snowflake, Redshift, BigQuery) and object storage (S3), with SQL queries executed server-side on the warehouse infrastructure rather than pulling data into Hex. The system supports parameterized queries where input controls (filters, date ranges) are passed to the warehouse, enabling efficient filtering at the source. Large datasets are handled through 'query mode' (details not specified) rather than loading entire tables into memory.

The Threads Agent (Team+ feature) enables multi-turn conversations where users ask follow-up questions and the agent maintains context across the conversation. Unlike the Notebook Agent which generates code cells, the Threads Agent appears to provide conversational analysis and insights. The exact implementation (whether it generates cells, maintains separate conversation state, or uses RAG over notebook history) is not documented.