EarCue

An Android prompter that reads your speech into your earphones, one line per thumb press.

Download APK · Script format · Issues

English | 繁體中文 | 简体中文 | Español | Português (Brasil) | 日本語 | 한국어 | Français | Deutsch | Русский

Features

• Instant switching: every press of Next or Previous cuts the current line and starts the new one, so lines never overlap. Hold to skip line by line, hold Play to restart from the first line, and turn on chimes for a new slide or a change of speaker.
• Earphone guard: with no earphones connected, or with media volume at 0, play buttons vibrate three times and make no sound. Unplugging mid-line stops playback.
• One-thumb control: three preset button layouts, drag-to-move and drag-to-resize editing with reset, a left/right-hand mirror, and a black Stage mode for a phone held behind your back.
• Hardware buttons: Sony shutter button (press = next, hold = keep skipping), optional volume keys, and wired or Bluetooth earphone buttons.
• Voices: Android text-to-speech with speed 0.25×–2.00×, a female/male switch, six preset voices, a separate voice and pitch for partner lines, word-by-word highlighting, or your own recording per line.
• Script library: local SQLite storage with search over titles, tags and text, tag filters, and a git-like history: update a script from a new file and restore any earlier revision.
• AI conversion skill: skills/earcue-script turns a Word, PDF or slide transcript into a tagged .earcue.json file and checks it word for word against the source.
• 10 UI languages: with light and dark themes: English, 繁體中文, 简体中文, Español, Português (Brasil), 日本語, 한국어, Français, Deutsch, Русский.

Demo

Try it with the fictional sample talk in examples/green-streets/: a slide deck, a bilingual transcript, and the converted cue file for presenter P2.

Architecture

flowchart LR
  T["Transcript<br/>.docx / .pdf / .pptx"] -->|earcue-script skill| J[".earcue.json"]
  J -->|Import file · Paste · Share| P[ScriptParser]
  P --> DB[("SQLite<br/>scripts · revisions · cues · recordings")]
  DB --> E[CueEngine]
  K["Touch deck · Sony shutter<br/>volume keys · earphone buttons"] --> E
  E --> G{"Earphones connected<br/>and volume > 0?"}
  G -->|yes| S["TextToSpeech or own recording"] --> H((Earphones))
  G -->|no| V["Vibrate ×3, no sound"]

Loading

Every import becomes a revision of a script, and the current revision is expanded into cues. CueEngine owns playback: it flushes the previous utterance before starting the next one and checks the audio output on every press. Details: docs/architecture.md

Technical details

Tech stack

Part	Choice
Language	Kotlin 2.1, JDK 17
UI	Jetpack Compose (BOM 2024.12.01), Material 3, custom drawn icons
Storage	Platform SQLiteOpenHelper (no ORM), SharedPreferences for settings
Speech	TextToSpeech, MediaPlayer for own recordings, MediaSession for earphone buttons
Build	Android Gradle Plugin 8.7.3, Gradle 8.11.1, compileSdk 35, minSdk 26
Release APK	about 1.4 MB, minified with R8

Script format

{
  "format": "earcue/1",
  "title": "Green Streets at Night · P2 Background",
  "speaker": "P2 Mina Park",
  "lang": "en-US",
  "tags": ["demo", "paper-review"],
  "sections": [
    {
      "title": "Why are summer nights getting hotter?",
      "slide": "P2",
      "minutes": 2.5,
      "cues": [
        { "partner": "P1 Alex Rivera", "text": "Quick poll first. ..." },
        { "text": "I am one of them. ...", "note": "On-screen hint" }
      ]
    }
  ]
}

A plain-text format for pasting is also accepted. Full specification: format.md

Permissions

Permission	Why
VIBRATE	Haptic feedback and the earphone-guard warning
RECORD_AUDIO	Only when you record your own voice for a line; asked at that moment

The app declares no internet permission. Online TTS voices are fetched by the speech engine, not by EarCue.

Installation

Download EarCue-<version>.apk from Releases, open it on the phone and allow installing from that source. Android 8.0 or later is required.

To build from source (JDK 17 and Android SDK 35):

./gradlew assembleRelease

Full guide, including the AI skill setup: docs/installation.md

Project Structure

EarCue/
├── app/src/main/
│   ├── java/app/earcue/   # audio/ data/ ui/
│   └── res/               # icons, strings × 10 languages (generated)
├── skills/earcue-script/  # AI conversion skill
├── examples/green-streets/# fictional demo
├── tools/
│   ├── i18n/              # UI string tables + generator
│   ├── readme/            # README template + translations
│   ├── audio/             # reminder tone generator
│   └── demo/              # demo file builder
├── design/                # logo source
└── docs/                  # installation, architecture, assets, README/ (translations)

Contributing

Issues and pull requests are welcome. UI text lives in tools/i18n/; run python tools/i18n/gen_strings.py after editing it. CONTRIBUTING.md

License

Free for personal, educational, research and non-profit use under MIT terms; commercial use needs the author's written permission. LICENSE