Privacy Policy

Two ways to run this

FounderVoice AI is open source, and where your audio goes depends entirely on which copy you are using. The hosted demo and a self-hosted install behave very differently, so both are described below.

The hosted demo

If you are reading this on our public demo site, treat it as a shared sandbox rather than a private workspace:

• Your recording is uploaded to our server, converted, analysed, and stored there along with its transcript, metrics, and coaching report.
• Audio is sent to Groq for speech-to-text. Their handling is governed by their own privacy terms, not ours.
• There are no accounts. Your recordings live in a private workspace identified by a cookie in your browser, so other visitors cannot see or delete them. Clearing that cookie, or opening the site in another browser, starts a new empty workspace and makes the old one unreachable.

Because there is no password, treat this as convenient rather than secure. It suits practice material and interview answers. Do not upload unreleased financials, customer information, or anything else you would not want stored on a server you do not control.

Running it yourself

When you run FounderVoice AI on your own machine or server, microphone audio, recordings, transcripts, and the SQLite database stay on that machine. Set ASR_PROVIDER=local and transcription runs in-process through Whisper, so no audio is sent anywhere. This is the configuration the project is designed around.

Optional cloud coaching text

If a coaching API key is configured on the API server, derived text and metrics (not raw audio) may be sent to that provider to generate coach language. Leave the key unset to use the built-in templates instead.

The coach’s spoken voice

When the coach reads your review aloud, the speaking happens in your own browser using the voice already installed on your device. Nothing is sent anywhere to produce it, and your recording is never used to make the voice. An operator can optionally switch the API to a hosted speech provider, in which case the coaching text — never the audio — is sent there to be spoken.

Contact form

If you submit the in-app contact or feedback form, we receive the details you enter (name, email, phone, company, message) so we can reply. Those details are delivered to our inbox by FormSubmit, a form-to-email relay, and are also stored on the FounderVoice API server. Nothing you record is attached. Do not include sensitive audio or secrets.

No unnecessary tracking

The app does not embed advertising trackers, and analytics are not required to use the product.

Your control

Use Fresh start in Coach to wipe sessions and Voice Memory. On a self-hosted install, deleting the data/ folder removes every stored file.

Contact

Use Contact us in the app sidebar, or the repository maintainer email if published.