Developer & AI data

Make small fixtures and repeatable experiments, with the generation rules visible.

Tools in this category

15 tools

Password generator

Create random passwords, memorable passphrases and PINs with browser cryptographic randomness.

UUID generator

Create version 4 UUIDs for fixtures, prototypes and test records.

Random strings

Characters for your next test.

Mock data generator

Build typed test records with a JSON column schema, repeatable seed and up to 1,000 rows. Export CSV, JSON, JSONL and a generation recipe.

Sample Data Lab

Create seeded classification, regression or clustering datasets with known generating rules. Preview points and export data, separate answers and recipes.

Dataset fault injector

Create clean and damaged fixtures with missing values, duplicate IDs, type errors or outliers. Export both versions and an exact change manifest.

Split & sampling inspector

Explore random, stratified, grouped and chronological splits on generated fixtures. Inspect partition sizes, class counts and group overlap; export assignments.

Known-answer evaluation packs

Generate repeatable arithmetic, sorting, date, lookup and extraction test prompts. Export JSONL inputs and separate exact answers for LLM smoke tests.

Normal distribution generator

Generate Gaussian samples with a chosen mean, standard deviation and seed. Inspect a histogram and summary, then export CSV, JSON or JSONL.

Discrete distribution generator

Sample Bernoulli, binomial, Poisson or custom weighted outcomes. Set parameters and a seed, inspect frequencies and export every observation.

Probability simulator

Run seeded coin, die or weighted-outcome experiments. Compare observed counts with expected counts and export the trial sequence and report.

Random date and time

Pick date-times on a clear UTC timeline, ready for sample data.

Random dates

Pick a day within your range.

Random times

Choose a time, down to the minute.

Color palette generator

Find colors that work together. Keep favorites and try new combinations.

Choose data for the job you are testing

Use UUIDs, strings and UTC timestamps for individual example values. Mock data builds typed rows; Sample Data Lab creates controlled classification, regression and clustering examples. Fault injection supplies clean and damaged fixtures with a change record. Split inspection and known-answer packs help check specific pipeline behaviors. Research tools export data and versioned recipes. The password generator is separate: it uses cryptographic randomness and never puts generated passwords in saved history.

Can these datasets prove an AI model works in the real world?

They are small, controlled examples for teaching, debugging and checking assumptions or output formats. They do not establish real-world coverage or production model quality. The current research tools generate up to 1,000 rows locally; fault and split tools use their own generated fixtures. Export the recipe and separate answer key when offered, and keep exported data when exact archived values matter.