OpenViSTAI · Agents · Blockchain · Crypto
Menu
AI Developer Tools & SDKs

Budoux

Budoux is a standalone machine-learning tool that predicts phrase boundaries to organize line breaks in text. Its Python, JavaScript and Java implementations provide parsing and HTML-output workflows, with pretrained models for Japanese, Simplified Chinese, Traditional Chinese and Thai.

Independent OpenViST project profile, based on linked sources.

Budoux — OpenViST project profile in AI Developer Tools & SDKs.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Budoux when ordinary wrapping risks breaking Japanese, Chinese or Thai text at awkward points and a small standalone segmentation component suits the text-display workflow.

Who it is for

  • Developers building text-processing pipelines or interfaces that need phrase-aware line wrapping.
  • Teams working with Japanese, Chinese or Thai text display.
  • Developers who want to supply a custom model or explore training for another language.

Documented uses

  • Split supported-language text into phrases to guide line wrapping.
  • Add phrase-aware break opportunities to HTML output.
  • Format text into phrase segments from a terminal.
  • Use a custom JSON model with the parser.
Before you choose

Requirements & limits

The project is open source under the Apache-2.0 license. The README documents Python installation with pip and links package listings for Python, JavaScript and Java.

  • Pretrained models are documented for four languages: Japanese, Simplified Chinese, Traditional Chinese and Thai.
  • Implementations are listed for Python, JavaScript and Java; the detailed installation and usage examples in the README are for Python.
  • The project describes its size as around 15 KB including the machine-learning model and says it does not depend on third-party word segmenters.
  • Budoux does not sanitize HTML input; untrusted input requires a separate appropriate sanitizer.
  • For Korean, the README recommends considering CSS word-break: keep-all as a general alternative to installing Budoux.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install Budoux and parse text with Python

Install the Python package and use a pretrained Japanese parser to split a sentence into phrases.

Before you start

  • Python package installation with pip
  • A sentence in a supported pretrained-model language

Steps

  1. Install the package with the command below.

    Documented command
    pip install budoux
  2. Import budoux in Python.

  3. Load the default Japanese parser with budoux.load_default_japanese_parser().

  4. Pass a sentence to parser.parse() to obtain a list of phrases.

Expected outcome

A list of phrase segments that can be used to organize line wrapping.

Guide limitations

  • The example uses Japanese; pretrained models are also documented for Simplified Chinese, Traditional Chinese and Thai.
  • The guide does not cover training a custom model.
Source-based practical guide

Format text into line-break segments with the CLI

Use the documented command-line interface to format text with a selected supported language.

Before you start

  • Budoux installed and its budoux command available in the terminal
  • Text in Japanese, Simplified Chinese, Traditional Chinese or Thai

Steps

  1. Provide text to the budoux command.

  2. For a language other than the default Japanese, specify -l with zh-hans, zh-hant or th; use ja for Japanese.

  3. Read the formatted phrase segments in the command output.

Expected outcome

Text displayed as separate phrase segments for line-break organization.

Guide limitations

  • The documented CLI language options are Japanese, Simplified Chinese, Traditional Chinese and Thai.
  • The guide covers formatting text, not HTML sanitization.
Evidence and transparency

Sources & review

Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
google/budoux · Official homepage title · retained in its source language