sylabach

Sylabach is an open-source software project that provides automatic syllabification and phonotactic analysis for natural language text. It offers tools to segment words into syllables and to annotate syllable boundaries according to language-specific rules. The project aims to support researchers, educators, and developers working with speech technologies and language learning applications.

Developed at the Institute for Computational Linguistics at Northbridge University, Sylabach began as a research prototype in 2010 by a team led by Professor Elena Morozova. It reached a stable public release in 2012 and has since undergone several major revisions to expand language coverage and improve accuracy. The project is maintained by an international community of contributors under an open-source license.

Sylabach’s architecture combines rule-based hyphenation rules with probabilistic models to resolve ambiguous boundary cases. It supports a core set of languages and can be extended via language packs. The library offers APIs for Python and Rust, and can be integrated into text-to-speech pipelines and educational tools. It also includes a benchmarking suite and a collection of example corpora.

Core features include syllabification, stress prediction, and optional phonetic transcription. Language packs cover typologically diverse languages, including Romance, Germanic, Slavic, and some agglutinative languages, though coverage varies by pack. The tool can operate in batch mode or as a service within larger NLP workflows.

Scholarly reviews note Sylabach as a practical tool for syllable-aware processing, though accuracy is dependent on language data quality and may be lower for languages with complex syllable structures or lacking comprehensive packs. Some critics call for improved handling of loanword adaptation and proper nouns.

Licence: MIT License. See also: Syllable, Phonotactics, Text-to-speech.