Milestone 1 Complete

The Swiss knife of phonetics.

A high-precision, offline-first mobile application tailored for linguists, English phoneticians, accent coaches, and language learners. OpenSchwa bridges academic phonological rigor with fluid, modern mobile design.

Academic phonological rigor meets offline speed. 146,000+ words with simultaneous RP and GA phonetic entries, context-aware weak form reductions, linking [r], and dual Gimson/Upton notation — computed 100% locally on your device.

Get the latest build View the source

What it does

Academic phonology tools and dialect-adaptive transcription right in your pocket:

Precision IPA Keyboard

Dialect-adaptive layouts switching instantly between RP and GA. Dedicated rhotic monophthongs ([ɝ], [ɚ]), allophones ([ɾ], [ɫ], syllabic consonants), and strict distinction between narrow phonetic [...] and phonemic /.../ boundaries.

146,000+ Offline Lexicon

Zero-latency bundled database covering 146,545 words with simultaneous RP and GA entries aligned to J.C. Wells’ (1982) lexical sets (THOUGHT, LOT, BATH, CLOTH), plus intelligent rule-based G2P fallback.

Connected Speech Engine

Automatic contextual reduction of unstressed function words (to, the, you, auxiliaries) and automatic junction linking [r] in non-rhotic RP (four apples → [fɔːr ˈæp.l̩z]).

Dual Notation System

Toggle dynamically between Classic British EFL notation (Gimson / Jones: /e/, /eə/) and Contemporary academic transcription (Oxford / Upton: /ɛ/, /ɛː/) across both the keyboard and transcriber.

Custom User Lexicon

Import custom JSON vocabulary overrides with priority lookup over the bundled database. Inspect active entries, modify phonetic keys, and export dictionary files directly within the app.

AI Verification (BYOK)

Bring-Your-Own-Key integration for DeepSeek, OpenAI, Groq, and OpenRouter. Delivers canonical IPA validation and concise allophonic feedback with zero external telemetry and local key storage.

How it works

Built with Flutter and Dart, OpenSchwa delivers a responsive, distraction-free Material Design 3 interface with monochromatic styling and haptic touch feedback.

Everything runs on your device. The 146,000+ word lexicon, phonological reduction rules, and G2P approximation are bundled locally for zero-latency offline execution. Your transcriptions and data never leave your phone, with AI verification connecting directly to your chosen provider using your own API key.

Roadmap

M1 done — Core Foundation & Transcription: MD3 UI & IPA keyboard, 146k offline lexicon (RP & GA), connected speech engine (weak forms & linking [r]), dual notation system, custom JSON lexicon, and BYOK AI verification
M2 — Acoustic Engine & Pitch (F₀) Visualization: real-time microphone input, low-latency audio capture, YIN / McLeod Pitch Method (MPM) F₀ tracking, and interactive intonation contours
M3 — Formant Analysis & Voice Onset Time (VOT): F₁ vs F₂ vowel space chart mapping, spectrogram rendering, and plosive VOT measurement

Releases

Downloads are published on GitHub Releases. Milestone 1 is complete — grab pre-built release APK binaries from GitHub, or compile the app for Android and iOS directly from source with flutter build apk --release.

Go to GitHub Releases ↗