Random Generator now includes 124 built-in generators for names, characters, places, science-fiction settings, usernames, pets, and more. The update also makes it easier to explore the catalog and build custom generators.
Find the right generator quickly
The interactive showcase lets you search and filter by category, choose formats and supported types, and generate batches of results. Set a seed to replay a sequence, copy results to the clipboard, and inspect an API example for the selected generator. Definitions load only when selected, so browsing the catalog doesn’t eagerly build every generator.
When the built-in data isn’t the right fit, the Custom Training Lab lets you experiment with Item and Markov generation.
Tune formatting and generated length
Output now uses title casing by default, with per-token overrides for titleCase, capitalize, uppercase, lowercase, and preserve. Markov generation supports minimum and maximum lengths, including per-pool minimums, and format patterns can be selected by weight.
The update also corrects curated pools, restores missing patterns, and improves seeded randomness, supported-type selection, and duplicate weighting. Shared models, compiled Markov transitions, cached casing, and reusable buffers reduce repeated work; Prewarm() lets you compile during loading without consuming random draws.
One compatibility note: output casing and seeded sequences may differ from older releases. The documentation covers the updated behavior and the full generator list.