Live: paua-ripple.shunpy.net

Browser-based voice dictation powered by the Aqua Voice Avalon API.

Unofficial BYOK client — your API keys stay in your browser and are sent directly to Aqua Voice and OpenAI. This site never proxies your keys.

• Record your voice with a push-to-talk or toggle button
• Transcribe via the Aqua Voice Avalon API directly from the browser
• Insert dictation at the cursor position in an editable transcript
• Optionally select a span of text and dictate an edit instruction — OpenAI rewrites the selection in place
• Live mic waveform and dB meter while recording
• History pane to restore any previous transcript state

1. Go to app.aquavoice.com/api-dashboard
2. Sign in or create an account
3. Generate an API key

Aqua Voice Avalon API pricing: $0.39 per hour of audio.

1. Go to platform.openai.com/api-keys
2. Sign in or create an account
3. Create a new secret key

OpenAI pricing: the default model (gpt-4.1-nano) is very cheap for edit use.

npm install
npm run dev

Open http://localhost:3000.
Click ⚙️ Settings and enter your API keys.

npm run build
# Output is in out/ — deploy anywhere that can serve static files

Suitable for GitHub Pages, nginx, Apache, or any static file host.

1. Open ⚙️ Settings and enter your Aqua Voice Avalon API key
2. Choose a language (Auto-detect / English / Japanese)
3. Record:
   • Hold to talk — hold the blue button while speaking, release to transcribe
   • Toggle record — click the teal button to start, click again to stop
4. Transcript appears in the textarea and is automatically copied to clipboard
   • If textarea has a cursor, dictation inserts at that position
   • If textarea is empty or unfocused, dictation replaces the content
5. To AI-edit: select a span of text, then record an edit instruction

All settings are stored in your browser only (localStorage).

If PauaRipple is useful to you, consider buying me a coffee:

MIT — see LICENSE.