Local speech-to-IPA recognition and vocabulary matching with lexicon-constrained phrase decoding.
Stop embarrassing names before they launch. Curser is an interactive application that takes in spoken or typed inputs and finds the closest curse word in multiple languages to the input.
Phonetic analyser of native English speech: time-aligns what was actually pronounced (wav2vec2 phone CTC) against the canonical pronunciation (forced alignment on the same acoustic pass), labels connected-speech phenomena (flapping, t-deletion, vowel reduction, linking) and prosody. CLI + local web UI. Docs in Spanish.
Phonetic-first language phylogeny: unified datasets (PHOIBLE, WALS, Glottolog, Lexibank…), feature builders (PanPhon/CLTS), and UMAP/HDBSCAN clustering with evaluation.
🌍 Unify and analyze phonetic data to reveal language relationships with PhonoPhylo, a scalable tool for phonetic-first language phylogenetics.
To associate your repository with the panphon topic, visit your repo's landing page and select "manage topics."