Introducing datakumi

Every report is only as good as the lists behind it: products, customers, suppliers, accounts and code lists. Today those lists live in spreadsheets nobody owns. datakumi gives each one an owner, rules and an approval step.

The problem

There are several copies of the product list, and none of them is current. Wrong codes are found after the report is already wrong. Approvals are buried in email threads. Data teams end up cleaning the same lists again and again.

What datakumi does

• Edit in a grid people already know. Changes stay in a private draft until they are published.
• Check every value against rules: keys, required fields, allowed values and duplicates.
• Approve changes with the right people, from the app, email, Teams or Slack. Nobody approves their own change.
• Publish the result as a Delta table in your own storage, with every version kept.

Your cloud, your data

datakumi runs as one container with a PostgreSQL database and your object storage: Amazon S3, Azure or Google Cloud. Your data stays in open formats, so Spark, Databricks, dbt, DuckDB and pandas read it directly. There are no export jobs and no lock-in.

How to get it

Self-hosting is free. If you would rather not run it yourself, the managed plan runs datakumi for you on your own data lake. See pricing, read the docs, or book a demo and bring one spreadsheet.