Your browser. Your voice studio.

Type it. Hear it.

Paste your script, choose a voice, and download your audio. No fixed word limit. No account needed.

1Paste text2Choose a voice3Generate & download

1Your script

0 characters · No fixed limit

0Words

0:00Estimated duration

155Estimated words / min

Michael · American maleChange voice

Preview checks the first 35 words. Generate records your whole script. Keep this tab open; completed parts are saved for recovery.

Automatic acceleration · Opus 64 kbps output

Use your NVIDIA GPU directly

One local NVIDIA helper powers both Audio and Studio. Once paired, both pages reconnect after refresh while the helper is running. Audio speed depends on your GPU; 10× is a target, not a guarantee.

Install Python 3.11 for Windows with the Python launcher enabled.
Download / update NVIDIA helper Run it once. The same helper powers Audio and Studio, and future refreshes reconnect automatically. First setup downloads several GB; later launches reuse them.
Keep its window open. It opens this site when ready. Click Connect NVIDIA below and allow Local network access if your browser asks.

Check Torch copies

Start the helper to receive your connection link.

Ready

Recent scripts

Saved on this device

2Best recommendations

Seven starting picks. Tap a voice, then Preview voice to hear your own script. Rankings are starting suggestions, not an objective score.

Speed1.00×

Opus · 64 kbps by default · approximately 43–58 MB for 90–120 minutes.

More settings

Voice Download format Compressed audio quality

Opus uses variable bitrate. For editing software that does not accept Opus, choose WAV or MP3.

90–120 minutes: about 43–58 MB.

Generation options

Acceleration

The GPU model downloads about 330 MB once. CPU uses a smaller model. Both are cached when browser storage is available.

Pronunciation fixes

One correction per line: word = spoken spelling

English pronunciation is improved automatically for all voices. Optional corrections for names take priority without changing your script. Preview a short sample before a long recording.

Runs locally in English, without an API key. First use downloads the selected voice model and runtime. After that, the engine stays ready for your next recording.

Volume100%

Instant device preview

Uses your device's own voice. This preview cannot be downloaded and sounds different from the selected studio voice.

Device voice

Pitch1.00

Pitch affects device preview only. These are original studio voices.

Sound effects

Standalone previews, not mixed into your recording.

Type it. Hear it.

1Your script

Your recording

Saved on this device