How are the weights distributed?

Weights follow a Zipf-style decay where each word's base weight is the maximum divided by its rank. A small random jitter is applied so reruns vary, and weights are clamped to a minimum of one.

What key names does the output use?

You can choose between text and value, which suits many JavaScript word cloud libraries, or word and weight, which other tools expect. Pick whichever matches your renderer to avoid remapping.

Can I supply my own words?

Yes. Select the custom topic and paste words separated by commas or new lines. They are trimmed, blanks removed, and then weighted with the same Zipf decay as the built-in topics.

Why is the first word so much larger than the rest?

Because weight is inversely proportional to rank, the top word gets the full maximum weight while the second gets half, the third a third, and so on. This mirrors how natural-language frequency typically looks.

Is the data reproducible?

The generator is seeded, so the same topic, count, and seed produce the same list. The Regenerate button advances the seed, changing the jitter and word selection.

Word Cloud Data Generator

A word cloud data generator that produces a frequency-weighted word list shaped for word cloud and tag cloud libraries. It outputs JSON with either text and value or word and weight keys, ready to drop straight into your renderer.

How it works

Choose a built-in topic — tech, marketing, finance, or health — or switch to custom mode and paste your own words. The generator picks the requested number of words and assigns each a weight using a Zipf-style decay: the top word receives your chosen maximum weight, the second receives half of it, the third a third, and so on by rank. A small seeded jitter perturbs the weights so reruns differ, and every weight is clamped to at least one.

You select the key names to match your library, and the output is reproducible because it is seeded. The Regenerate button advances the seed, changing both the word selection and the jitter.

Tips and example

A max weight of 100 with 15 words gives a natural-looking spread for most renderers. The output looks like this:

[
  { "text": "cloud", "value": 96 },
  { "text": "data", "value": 52 },
  { "text": "API", "value": 31 }
]

If your library expects word and weight, switch the key-name dropdown rather than remapping by hand. For custom topics, paste a focused word list — overly long lists dilute the Zipf curve and the largest words stop standing out.