About
A spreadsheet that thinks with you.
Analysts spend half their week moving data. We're building the tab where it just works.
Rows are sacred
We never alter source data. Every answer traces back to rows.
Fast is a feature
A million rows should feel like a hundred.
Show the work
AI answers ship with the query that produced them.
How we test
Tested across 10 industries before every release
Every change runs 120 automated checks: ten synthetic industry workbooks, each asked twelve kinds of question. Expected answers are worked out separately, never by the engine being tested. We use made-up data on purpose — we don't test on customer data, and customer data doesn't train our models.
- 01Totals and filtered sums
- 02Rates by group
- 03Combined-category shares
- 04Per-row thresholds
- 05Weighted averages
- 06What-if recalculation
- 07Period-over-period growth
- 08Weekday vs weekend
- 09Ranking within groups
- 10Distinct counts with duplicate IDs
- 11Zero and blank denominators
- 12Question-type routing
A new name, built on years of experience
Creeklytics is made by CreekData. We are just getting started with our own brand, but the people behind it have spent years working hands-on with data. We built the tool we always wanted: one that handles huge files, works with company data where it lives, and never makes up a number.
Data engineering
Years building pipelines, warehouses and lakehouses that companies run on every day.
Data analytics
We have sat in the month-end crunch and answered the same questions again and again.
Data science & ML
Forecasts, driver analysis and statistics that hold up when someone checks the maths.
Language AI (NLP)
Turning plain-English questions into exact calculations instead of guesses.
Data professionals from teams at Microsoft, AWS, Apple and Jio are exploring Creeklytics early and helping us improve it with honest feedback. Their company names show where they work, not an endorsement or a customer agreement.