Privacy

Short version: your audio stays on your device unless you press a button that clearly sends it somewhere.

Stays in your browser

The recorder, mic test, editor, effects, trimmer, joiner, converter, voice changer, pitch and speed tools, noise remover and vocal remover all run locally using your browser's Web Audio API. Audio is held in memory and discarded when you close or reload the tab. It is never uploaded.

Sent on request

• Audio to Text: when you press transcribe, a compressed 16 kHz mono copy (up to 10 minutes) is sent through Vercel AI Gateway to Google's speech-to-text model. If you ask for show notes or a summary, the transcript text is sent to an OpenAI language model the same way.
• Text to Speech: the text you type is sent through Vercel AI Gateway to Google's text-to-speech model.

VocalForge does not store this audio or text. The providers process it to return a result under their API terms. Your IP address is held briefly in server memory to enforce hourly usage limits.

Saved on this device

We use your browser's local storage for your preferred export format and bitrate, recorder processing options, and your last 10 transcripts. None of it leaves your device.

Clear all VocalForge data on this device

Analytics

We use Ahrefs Web Analytics, which counts page views without cookies and without building personal profiles.

Consent

Only record, transcribe or transform voices of people who have agreed to it. VocalForge has no voice cloning and does not identify speakers.