Word Cloud
How to Use
📖 Tool Introduction
A pile of survey responses, user comments or a long report rarely reveals its emphasis at a glance. A word cloud scales each term by how often it appears and separates them by color, so the frequent ones stand out immediately. The result drops neatly into a slide, a poster or an analysis page.
Segmentation and drawing both happen in the browser. Paste text and the tool counts frequencies first, then walks an outward spiral placing words one by one, sizing each by its count and skipping any that no longer fit. You can also paste a 'word,count' table and skip the counting step entirely.
Chinese segmentation works without a dictionary by looking at how often adjacent characters co-occur, backed by a stop-word list that drops low-information fillers. This approach handles domain vocabulary well because specialist terms are not chopped up by a fixed lexicon.
Everything runs locally, so research transcripts and customer feedback never leave your machine.
✨ Features
- Automatic Chinese frequency counting via character co-occurrence, with no fixed dictionary required.
- Accepts a pasted 'word,count' table to bypass automatic segmentation.
- Built-in stop-word filter removes low-information fillers with one click.
- Five palettes: blue-violet, warm orange, fresh green, pink-purple and deep navy.
- Font size follows frequency, or can be flattened to pack more words in.
- Rectangle or ellipse outline, with horizontal, vertical or mixed text direction.
- Canvas size, word cap and font range are all adjustable, from avatar thumbnails to poster sizes.
- Lists the top thirty terms by frequency so you can spot segmentation mistakes.
- Reports how many words were actually placed and warns when the canvas runs out of room.
- Exports PNG and supports one-click copy to clipboard.
- Fully client-side; research text and feedback are never uploaded.
🚀 How to Use
- Paste your article, comments or survey text, or upload a .txt file.
- If you already have counts, paste one 'word,count' pair per line instead.
- Set canvas size, maximum words and the font size range, then pick a palette.
- Optionally enable stop-word filtering and frequency scaling, then press generate.
- Review the preview and ranking, then download the PNG or copy it to the clipboard.
⚠️ Notes
- Segmentation is statistical：Without a dictionary, phrases can be split or merged unexpectedly. Supply a frequency table for proper nouns you care about.
- Filtering can over-remove：Turning the filter off keeps more words but adds noise; for specialist text, compare both results before deciding.
- Small canvases run out of room：Many words at large sizes may not fit. The tool reports the placed count, so enlarge the canvas or reduce the word cap.
- Single characters need a tweak：The default minimum length is two. Set it to one to count single characters, at the cost of noticeably more noise.
- Fonts vary by system：Rendering uses system fonts, so output differs slightly across devices. For final artwork, export and re-typeset in a design tool.
- Colors cycle by order：Palette colors cycle through the word order rather than mapping to frequency, so equal-frequency words may differ in color. This is intentional.
Input and word handling
- Stop words: a built in list filters common Chinese and English function words so that rare, meaningful terms dominate the picture.
- Grouping: multi word phrases stay intact when you enable phrase mode, which keeps product names and slogans from being split apart.
Privacy and compatibility
- Nothing is uploaded: tokenising and layout run in your browser, so survey answers and internal documents never leave your device.
- Readability: reduce the word count or raise the minimum frequency when the layout becomes crowded, and export PNG for slides or SVG for further editing.