Introduction

What is Codatum?

Codatum is a data analytics platform for connecting to a data warehouse (DWH), running SQL, and analyzing, visualizing, and sharing the results. It supports BigQuery, Snowflake, Databricks, and Redshift.

You do your analysis in a notebook. Because SQL, charts, and text stay together in a single document, your team can share not just the numbers but also the logic behind them and how to interpret them.

An AI agent can operate directly on your notebooks. From a natural language instruction, it runs SQL and creates charts, leaving the process as SQL in the notebook. It draws on descriptions and tags in the catalog and definitions accumulated in agent context to inform its analysis.

What you can do with Codatum

Analyze

Write SQL in a notebook and bring the results together with charts and text in a single document. SQL blocks support autocomplete and formatting, and you can build up SQL step by step by referencing SQL defined in other blocks (SQL chaining). Parameters let you change conditions like a date range or analysis target with ease.

For standardized analyses such as funnel analysis or budget-vs-actual tracking, you can arrange charts on a grid page to build a dashboard. For deep dives into changes in the numbers, use the explorer to explore results through a GUI.

Delegate to the AI agent

Ask the AI agent a question in natural language, and it runs SQL and creates charts for you. Because the process is left as SQL in the notebook, you can review and adjust the output yourself.

The agent draws on descriptions and tags added in the catalog and metric definitions placed in agent context, enabling analysis that reflects internal terminology and calculation rules. You manage the model, permissions, and fixed context in an AI profile.

Share

You can choose how to share your analysis results depending on the audience and purpose. Use a report for internal members and guests, a public link when you want to show results without requiring login, and a signed embed to embed results in your own SaaS product or internal tools.

Reports can be viewed and re-run within the scope of the publish settings even without connection permissions. With signed embeds, you can use fixed values for server-side parameters to isolate the data each tenant can access.

Run on a schedule

Use a workflow to run queries or reports on a schedule. You can connect the results to follow-up steps, such as delivering them to Slack or email, or summarizing them with an LLM. This is useful for things like daily KPI delivery or model monitoring.

Organize and manage your data

Register a connection and configure which tables to sync, and the table and column metadata syncs to the catalog. Adding annotations (descriptions and tags) makes that metadata available for autocomplete, search, and the agent.

Permissions are controlled through a combination of workspace roles and resource roles. Codatum also supports dividing work areas with teamspaces, granting permissions by group, audit logs, and IP access control and MFA.

Operate from a terminal or CI

With the Codatum CLI, you can treat notebooks as markdown files and catalog annotations as YAML files. This enables version control with git and operation from CI or coding agents.

Documentation structure

• Guide — Screen layout, global navigation, and more
• Features — Explanations of the main features
  • Quick search — Cross-search across notebooks, tables, and queries
  • Connection — Data warehouse connection setup and metadata management
  • Catalog — Metadata management for tables and saved queries
  • Query — Saved SQL reused within a connection
  • Notebook — Your workspace for data analysis, with collaborative editing, SQL, charts, and more
  • Workflow — Scheduled runs and automation
  • AI agent — Natural language analysis and notebook operation, and AI profiles
  • AI Assistant — A chat for creating SQL and charts and answering questions (legacy feature)
  • Codatum CLI — A CLI for operating Codatum from the terminal
• Workspace — The management unit, users, invitations, API keys, and more
• Permissions — The permission model and permissions for each resource
• Account — Accounts, sign-up, and account settings
• Reference — Glossary, Limits, System requirements