Straight answers about fake data.

The basics

New to this? Start here.

What is this for?

Demoing with real customer records is a bad idea. Demoing with data from 2019 is worse. This gives you invented customers, orders, sign-ups and amounts that look and behave like the real thing, so you can demo a product, fill a dashboard, test a spreadsheet, or hand a designer a screen that looks lived in.

Why do the dates not go stale?

Because they were never calendar dates in the first place. You say when things happened relative to a single reference day: signed up three months ago, ordered last week. Point that day at “today, whenever I generate” and every file you pull is current, whether that is tomorrow or next spring.

Will I get different data every time?

No, and that is the whole trick. The same model gives you the same rows every time, the same customers, the same orders. Only the dates move. So you can rehearse a demo, record a video, or point at one customer by name and know they will still be there next month.

Do I need to know how to code?

No, and there is nothing to install. Add a table, then pick what each column holds: a name, an email address, a date in a range you set, a number, a category with your own values. Sample rows appear as you work, so you are looking at the real result rather than guessing at it.

What do I get at the end?

A file you can hand to anybody, in one of four formats. XLSX is the usual pick: a workbook with one tab per table, which Excel, Google Sheets and Numbers all open. JSON puts every table in one file, for whoever is building the software. CSV gives you a single table, which is what most systems ask for when you upload into them. And SQL gives you every table as insert statements, so a developer can load the whole thing into a database in one go.

What you can trust

The fair questions to ask before using something new.

Is this data really invented, and can I show it to a customer?

Yes to both. No personal record is ever copied, sampled or anonymized from a real one. Names are assembled from common first and last names. Locations use real information for cities, regions, and postal prefixes, since that's what makes addresses look real, but house numbers and street names are always invented. Any resemblance to a real person or address is a coincidence. Email addresses use example.com, a domain reserved so it can never receive mail, and phone numbers come from ranges reserved for fiction, so they cannot be dialed.

How can I use the data?

However you like! Put it in a demo, a screenshot, a course, a pitch deck or your own product, at no cost and with no attribution, and you do not need to ask us. The one thing worth doing is labeling it as sample data anywhere somebody might take it for real results.

Does my information leave my computer?

The data you generate never does. Rows are built inside your browser and downloaded straight from the page, so they are never uploaded anywhere, not even to us. What reaches us is what you hand over on purpose: your sign-in details, any model you choose to save, and anything you write in a feedback message. We also count basic usage, like how often a download fails. Those records carry no account identifier, though one may note which model was involved. The privacy policy spells out exactly what. A saved model holds your own words, table and field names included, so keep real personal information out of them.

Can I save my models?

Yes. A free account gives you somewhere to keep them, and you can open, change, or delete a saved model whenever you like. Deleting is permanent. Note that we are in beta, so while we do our best to avoid it, a change on our end could occasionally affect your saved model. Either way, keep a downloaded file of anything you would hate to rebuild, and know that the file holds the data rather than the setup that made it.

It says free. What is the catch?

There is nothing to buy yet. Building and previewing are free and do not need an account, and the free account exists so you can save your models and download your data. Neither ever asks for a card. Pro is coming for people who want more models, more rows and a command line tool, but you can download as often as you like, on any plan. We do not sell your data and we do not run ads. More on the pricing page.

How it compares

Where this fits, and what is on the way.

How is this different from other data generators?

Most of them hand you a snapshot. The dates freeze the second you export, so a file that looked current in March looks abandoned by June and somebody rebuilds it from scratch. The rebuild is the real cost, because generating again somewhere else gives you a completely different set of people. Here it is the same set every time: same customers, same orders, same amounts, with only the dates moved forward. So regenerating gets you the data you already know, later, rather than a stranger's. And the tables agree with each other, so an order can never be dated before the customer who placed it.

Why not just ask a chatbot?

For twenty rows you will look at once, do. Past that, two things go wrong: you get a different answer every time you ask, so the data you rehearsed with is not the data you show, and the details quietly stop agreeing with each other. Where a chatbot wins is wording, which is why AI-written product names and review text are on the way here.

Who is this for?

Anyone who shows software to other people: sales engineers, founders on a pitch, teams filling a sandbox, designers who need a screen that looks lived in. What they share is using the same data more than once, which is exactly where a throwaway list of names stops being enough.

What is coming next?

That is where you come in! For now we are planning AI-written text for description and comment columns, a command line tool so you can generate from a script instead of the browser, and sharing a model with others. Tell us what you would like to see, though. What people ask for is what gets built next, and there is a link at the bottom of this page for exactly that.

Enough reading. Go build one.

Building a model and previewing your rows is free, and takes about two minutes.

Something here unclear, or something missing? That is our fault. Tell us.