~/fake-data ☀ LIGHT lorem sql api tester ☕ Support me apps ← about me
NAME

Fake Data Generator

— generate seeded test data: 131 field types, 18 sector schemas, tables that join, a CSV anonymiser
SYNOPSISfake-data [csv|xlsx|json|sql|ndjson|yaml|xml|md|html] schema
NOTESruns in this tab · nothing is uploaded · same seed + same schema = same data
0rows
0columns
—size
—time
—seed
id,name,email,status 1,Alice,alice@ex.com,active 2,Bob,bob@ex.com,pending …
Build a schema, then press Generate131 field types · 18 sector schemas · 9 locales · seeded

Guide — realistic test data in five steps

From an empty schema to a file you can import. In your browser, without an account.

1Start from a template

Open the Library tab and pick a template such as Person or E-commerce, or one of the 18 industry schemas. Or press Paste names and type your own column names, or paste a sample row, and the types are guessed for you.

2Shape the columns

Press Add column to search 131 field types, each with a live sample. Rename, drag to reorder, set weighted values, a share of empty cells with ∅% and U for unique values.

3Set rows, locale and seed

In Settings choose up to 50,000 rows, one of 9 locales and an export format. Type a seed and the same schema gives the same data on every run.

4Check the result

The Data view updates as you edit. Sort, filter and open a row. The Profile view shows nulls, duplicates, ranges and whether every checksum passes.

5Download or take it further

Download as Excel, CSV, JSON, SQL and more. Related makes tables that join, Anonymise masks a CSV you already have and Code turns the schema into SQL, Prisma, Zod or Faker.

Quick reference

PATTERNA mask for ids and codes
#
a digit 0-9
9
a digit 1-9
A / a
an upper / lower case letter
*
a letter or a digit
x
a hex digit
[abc]
one of the characters in the brackets
\
the next character, literally
Example
AA-####-[XYZ] gives QK-4817-Y
FORMULABuild a value from the same row
{column}
the value of an earlier column
{row}
the row number, from 1
|lower |upper
change the case
|strip |slug
remove accents, or make a url slug
|title |first
capitalise, or take the first word
Example
{first_name|strip}.{last_name|strip}@acme.be
COLUMNPer-column controls
∅%
the share of empty cells
U
force unique values
Weighted
active:70,inactive:20,banned:10
Grip
drag to reorder, or use the arrows
Type
press the type name to change it
KEYSShortcuts
Ctrl+Enter
generate
Ctrl+S
download
Ctrl+Z
undo a schema change
F
full screen
A
add a column
/
filter the rows
1 to 5
Data, Related, Anonymise, Code, Profile
Esc
close a window or leave full screen

What is a fake data generator?

A fake data generator, also called a test, mock, dummy, sample or synthetic data generator, creates realistic but entirely fictional records that you can drop into a database, an API mock, a spreadsheet or a design mock-up. You build a schema of the columns you need, choose how many rows you want and download the result. Everything on this page runs 100% in your browser: no upload, no account, no row-count paywall and nothing that could leak, which makes it a safe answer to “can we just copy production into staging?”.

131 field types

Identity, contact, location, business, finance, commerce, network, date & time, technical, text and special, from IBAN and EAN-13 to Kubernetes pod names and Belgian company numbers.

Valid checksums

Credit cards pass Luhn, IBANs carry a real mod-97 check digit, national and company numbers pass their own tests, EAN-13 and ISBN-13 carry a correct GS1 digit. Your validators will accept them.

Coherent rows

The email matches the name in the same row, the work email matches the company, the state code matches the state. Not a bag of unrelated values.

9 locales

US, UK, Belgium, the Netherlands, Germany, France, Spain, the Nordics or a global mix: names, cities, phone formats, postcodes, plates, IBAN country and national ids all follow.

Seeded and reproducible

Set a seed and the same schema produces byte-identical data on every run, on any machine. Safe to commit as a fixture and to use in snapshot tests.

Related tables

Parent and child tables whose foreign keys point at rows that exist, as one SQL file with CREATE TABLE and REFERENCES, as JSON, as a ZIP of CSV files or as an Excel workbook.

Anonymise your own CSV

Paste a real export and get a shareable one back: consistent, format-preserving masking with Luhn and mod-97 recomputed, and a leak check. Nothing is uploaded.

Schema as code

The same schema as PostgreSQL, MySQL, Prisma, TypeScript, Zod, Mongoose, Pydantic, SQLAlchemy, Go, JSON Schema, OpenAPI, Faker.js or Python Faker.

11 download formats

Excel, CSV, Excel CSV, TSV, JSON, NDJSON, SQL with optional CREATE TABLE, YAML, XML, Markdown and HTML, up to 50,000 rows.

The studio: schema on the left, data on the right

The page is laid out like a small data tool. The left panel holds the schema, the library of templates and sector schemas, and the settings for rows, locale, seed and format. The right side shows the result: a table that stays fast at 50,000 rows because it only draws the rows you can see, with sorting on every column, a filter box and a row view that shows one record as a list. The preview follows your edits by itself up to 5,000 rows, and above that you press Generate. Every structural change, such as a deleted column or a replaced schema, can be undone with Ctrl+Z.

Building the schema is quick. Paste names turns a list of column names, or one sample row as JSON or CSV, into a schema and guesses the type of each column from its name and its value. Add column opens a picker with a search box and a live sample next to every one of the 131 types. Your schemas can be saved in the browser by name, copied as JSON, loaded again from JSON and shared as a link that rebuilds the schema, the seed and the format. On a phone the schema panel slides up as a sheet, and full screen gives the data the whole display.

Realistic means more than random

Most generators give you a grid of independent random values: an email that has nothing to do with the name beside it, a state code that contradicts the state, a credit card number that your own validation rejects. That data is fine for filling space and useless for finding bugs. Here, each row is built with a little context: the person’s name drives the email, the username, the initials and the social handle; the company drives the work email, the domain and the website; the locale drives the city, the street format, the postcode, the phone number, the plate, the national id and the IBAN country. What comes out looks like a real export, so it breaks your code in the same places a real export would.

Field types by group

National ids, company numbers and plates that pass their checks

Test data for European systems needs numbers that look right and also pass the check the system runs. The National ID type follows the locale: the Belgian rijksregisternummer with its birth date, sequence and mod-97 control digits, the Dutch BSN with the 11-test, the French NIR with its key, the Spanish DNI with its check letter and the Swedish personnummer with a Luhn digit. Enterprise Number gives the company registration number of the locale, such as the Belgian 0xxx.xxx.xxx with a valid mod-97 pair or a French SIREN with its Luhn digit, and VAT Number gives VAT numbers with the right country prefix and, for Belgium, France and Sweden, a valid check. License Plate uses the national format, such as the Belgian 1-ABC-123. All of them are fictional: they pass the arithmetic, they do not belong to anyone. The Profile view re-checks every column in your own browser and tells you how many values pass.

Patterns and formulas for the columns no type covers

When a column is specific to your system, build it instead of searching for a type. A Pattern column fills a mask: # is a digit, 9 a digit from 1 to 9, A and a a capital or lower case letter, * a letter or a digit, x a hex digit and [XYZ] one of the characters in the brackets. So INV-2026-##### gives invoice numbers and AA-####-[XYZ] gives codes such as QK-4817-Y. A Formula column builds a value from earlier columns of the same row, with {first_name|strip}.{last_name|strip}@acme.be for an email that matches the name, {row} for the row number and the modifiers |lower, |upper, |strip (removes accents), |slug, |title and |first.

For numbers that behave like measurements there are three more: a normal distribution with a mean and a spread, a random walk that drifts from a start value, and a time series that steps forward by a fixed interval with a little jitter. Together they make a believable sensor feed, a stock-like price line or a daily metric.

Why generate test data at all?

Seeding a fresh environment so it looks real. Mocking API responses before the backend exists. Filling a dashboard for a screenshot without exposing a single customer. Load-testing an import with fifty thousand rows. Proving that your CSV parser survives a comma inside a company name, that your table layout survives a Belgian address and that your report survives a null. All of it without the legal and operational risk of copying production data into a system with weaker controls, which is the most common finding in a data-protection review.

Built for edge cases, not just volume

Download as Excel, CSV, JSON, SQL, NDJSON, YAML, XML, Markdown or HTML

Generate up to 50,000 rows and download them in eleven formats. Excel writes a real .xlsx workbook with numbers stored as numbers, dates stored as dates, a frozen header row and an autofilter, so it opens ready to sort. CSV is RFC 4180-escaped; CSV Excel adds a byte-order mark and semicolons so a European Excel opens it in columns with accents intact; SQL can write a matching CREATE TABLE with inferred column types and UNIQUE and NOT NULL constraints, in batches of 500 rows per statement; NDJSON feeds log pipelines and bulk loaders; Markdown and HTML are for documentation and mock-ups. Switching format re-renders the same rows rather than drawing new ones, so you can download one dataset five ways.

Sector data sets: start from your industry, not a blank schema

Almost nobody wants “fake data” in the abstract. They want sample insurance claims, test patient records, dummy invoice lines or example meter readings: a file that looks like the one their system really ingests. The library holds 18 ready-made industry schemas on top of the twelve generic templates, each with the columns, code lists and status mixes that the domain uses: a claims file with reserves and an excess column, a CDR file with cell ids and call duration, a lab file with analytes, units and a reference range. Pick one, change what you need, generate.

Each link loads that schema and generates it, so you can bookmark or share the exact starting point. Add &rows=1000, &fmt=xlsx or &locale=be to the address to fix the row count, the download format and the locale too, which is handy in documentation, a ticket or a README.

One table of random rows is easy. What breaks most test fixtures is the second table: an orders file whose customer_id points at customers that do not exist, so every JOIN returns nothing and every report looks empty. The Related view generates a small relational data set: parents first, then children whose foreign keys are drawn only from parent keys that were created. Four sets are built in:

The distribution is deliberately skewed rather than uniform: a few customers have many orders and some have none. That is what production looks like, and it is what exposes the report that silently drops customers without rows and the pagination that assumes an even spread. Under the tables you see the proof, one line per foreign key with how many rows join, how many parents have no children and how many children the busiest parent has. Download the set as one SQL file with CREATE TABLE, PRIMARY KEY, REFERENCES and inserts, which loads into an empty Postgres, MySQL or SQLite database as it is, as one JSON object with an array per table, as a ZIP with one CSV per table or as one Excel workbook with a sheet per table.

Anonymise a CSV you already have

The other half of the problem is the file you are not allowed to share. You have a real export and you need it in a ticket, a bug report, a demo or a shared staging database. The Anonymise view reads the file in your browser, works out what each column holds and replaces every identifying value with a fake one of the same shape:

Once a file is analysed you can press Make a generator from this file. It builds a schema with the same columns, null rates, value ranges and category mixes, and no original values are copied, so you can generate any amount of similar data from a file you were only allowed to look at. Nothing is uploaded: the file never leaves the tab, which is the only way a masking step is also a compliance step. It is a pragmatic data-masking and pseudonymisation tool for test and development environments, not a formal anonymisation certificate. A file with one Belgian rheumatologist in it can still be re-identifiable from the columns you kept, so drop what you do not need.

Export the schema, not just the rows

You rarely need only data. You need the table to put it in, the model that reads it and the type that validates it. The Code view turns the schema you built into working code in thirteen shapes: PostgreSQL and MySQL DDL (with UNIQUE, NOT NULL, CHECK or ENUM from your value lists), Prisma, TypeScript, Zod, Mongoose, Pydantic, SQLAlchemy, a Go struct with json and db tags, JSON Schema 2020-12, an OpenAPI 3.1 component and a ready-to-run generator script for Faker.js and Python Faker, so the same schema can move into CI, where a browser tab cannot go.

Nullable columns follow the ∅% you set, unique columns become UNIQUE or @unique, weighted lists become enums or check constraints and numeric ranges become minimum and maximum or a Pydantic Field(ge=…, le=…). It is a starting point you will edit, not a migration to apply blind.

Profile and validate what you generated

Generated data is only useful if it is what you think it is. The Profile view reads the dataset and reports, per column, how many distinct values it holds, the real null rate, the numeric range or the string-length range, a small distribution, the three most common values with their share and, where the type has one, a checksum verdict. Cards are checked with Luhn, IBANs with mod-97, Belgian national and company numbers with their own mod-97 rule, the Dutch BSN with the 11-test, and EAN-13 and ISBN-13 with the GS1 weighted sum. It also counts duplicate rows and tells you which columns are unique in every row, so you know what can serve as a key before you import.

This is the part that answers “is this data valid?” without a round trip through your own validator, and it is how you catch a null rate you set to 5% that landed at 40% because a column was drawn from a short list.

Cross-sector recipes: the awkward combinations

Real test data rarely sits inside one theme. A few combinations that come up constantly, and how to build them here:

Free Mockaroo alternative: what is honestly different

Mockaroo is the tool most people compare this to. Its free tier is capped at 1,000 rows per download and its multi-table relational export sits behind a paid plan; it is a hosted service, so your schema and the generated file go through a server. This page is client-side: no account, no row-count paywall and no upload, with relational sets, Excel download, schema and code export, CSV anonymisation and profiling included. What a hosted tool still does better is the part a browser tab cannot reach: a REST API you can call from CI, regex-driven custom types and datasets far beyond what a tab can hold in memory.

How it compares, checked September 2026

 This pageMockaroo (free)generatedata.comFaker.js / Python Faker
Rows per export50,0001,000~500 in the browserunlimited
Account needednofor saving schemasnono
Data leaves your machinenevergenerated server-sideclient-sidenever
Related tables with working foreign keys4 built-in setspaid plansnoyou write it
Industry-specific schemas18 sectors + 12 templatesfield types, not schemasa handfulno
Anonymise a CSV you already haveyes, in-browsernonono
Export the schema as code13 targetsSQL onlySQL onlyit is code
Profile & checksum verdictyesnonono
Reproducible from a seedyesnonoyes
Valid Luhn / mod-97 / GS1 valuesyesyespartlyyes
Runs in CIno — it is a pagepaid APInoyes
Works offline once loadedyesnoyesyes

Free-tier limits and paid features change; check the vendor’s current page before you rely on a row in this table. The honest summary: for a file you need now, in a browser, a page like this one wins on limits and privacy. For generating data inside a test suite that runs on every commit, a library wins, which is why the Code view hands you the Faker.js and Python Faker version of whatever you built here.

Privacy: what stays in the browser

The generator, the anonymiser, the Excel writer and the profiler are all JavaScript that runs in the page. After the page has loaded it makes no network requests of its own: no row, no schema and no CSV you open is sent anywhere, which you can confirm in your browser’s network tab. The only outside requests are the web font and the cookie-free page-view analytics. Your current schema, your settings and the schemas you save by name are kept in this browser’s local storage so they are there next time, and you can empty them from your browser settings. Old share links, with the schema packed into the address, still open exactly as before.

Want to tweak the generated data by hand? Paste it into the free online notepad to format JSON, sort or dedupe CSV lines and find & replace with regex, then save it as a file.

Lore GenPlaceholder text for developers SQL PlaygroundQuery CSV and Excel with SQL API TesterCall and inspect REST APIs Regex TesterTest patterns live Env EditorCompare and edit .env files All free tools →Every app on jasperbernaers.com