google/budoux

Health v2 · maintenance only

Adoption not part of the score

1772 stars · 44 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

BudouX is a small, standalone, machine-learning-powered line break organizer that segments sentences into phrases for languages without word boundaries like Japanese, Chinese, and Thai. It ships lightweight pretrained models (~15 KB) with bindings for Python, JavaScript, and Java, and can process HTML input.

Use cases

• prevent awkward line breaks in Japanese text on the web
• segment Chinese sentences into phrases
• split Thai text into words without spaces
• add zero-width spaces to HTML for CJK line wrapping
• run lightweight word segmentation client-side without external APIs
• train a custom phrase segmentation model for another language

When to choose

• you need CJK/Thai text segmentation with no external dependencies
• bundle size matters, e.g. client-side web typography
• you want pretrained models plus HTML-aware processing

When to avoid

• you need full tokenization or POS tagging rather than phrase-level segmentation
• your language uses spaces between words (e.g. Korean, English)
• you need heavy-duty NLP pipelines with rich linguistic analysis

Facets

library · maturity active

nlp machine-learning parser web-development frontend python jvm browser cross-platform text-segmentation line-breaking cjk japanese chinese thai typography html-processing natural-language-processing nodejs

Member repositories

For agents

markdown · JSON · MCP: product_card(name="google/budoux")