Describe any scene — a stormy tavern, a moonlit forest, a space station airlock…
2
GPT-5 breaks it into 5 distinct environmental sound layers
3
ElevenLabs generates each as a 30-second seamless audio loop
4
All layers can play together — adjust volume on each independently
🔒 Your keys, your data. Undertone runs entirely in your browser.
Your API keys are stored locally and requests go directly to OpenAI and ElevenLabs —
never through any intermediary. No accounts, no tracking, no backend.
Generated audio tracks are cached in your browser's IndexedDB so revisiting a scene
costs nothing — no repeat API calls.
📤 Share as a single file. Once a soundscape is ready, hit
Export to download a self-contained HTML file with all audio embedded.
Send it to anyone — they can open it in a browser and play it with no install,
no keys, no internet connection required.