Confidentiality: no AI training, no voice cloning

• Your audio is never used to train AI models, ours or anyone else's.
• We never create voice clones, voice models or voice fingerprints from it.
• Your uploaded file is deleted as soon as processing finishes.
• Your enhanced file is deleted 7 days later, automatically.
• Nobody at Optivox listens to your audio.
• Your audio is never sold, licensed or shared for datasets or advertising.

We will never use your uploads, outputs or any data derived from them to train, fine-tune or evaluate machine learning models, large language models or speech synthesis systems.

We will never build voice clones, synthetic voices or voice fingerprints from your audio, and we will never attempt to identify speakers from it. We require the same no-training, no-voice-cloning restriction from the processing provider we use.

The enhancement itself is signal processing: levelling, noise reduction, de-essing, loudness normalisation. It analyses the waveform of the file in front of it and nothing else. It does not learn from your recording, and processing one file has no effect on any other.

These commitments survive termination of your account. If we ever intend to change them, existing users get advance notice, and the change never applies retroactively to audio already uploaded.

1. You upload the file over an encrypted connection into private storage. It is not publicly reachable, and it has no shareable URL.
2. It is passed to our audio processing infrastructure to run the enhancement you selected, and only for the duration of that one job.
3. The enhanced MP3 and WAV are written back to private storage, reachable only through short-lived links issued to your signed-in account.
4. Your original upload is deleted immediately once the job completes. The enhanced output is deleted automatically 7 days later.

After that window, all that remains is job metadata: filename, duration, timestamp, status and minutes charged. No audio content.

Processing is fully automated. No member of the Optivox team plays, reviews or samples customer audio as part of normal operation. The only exception is if you personally ask us to look at a specific file to diagnose a problem you've reported. We will only ever do that on your explicit request, for that file.

If you would rather not wait for the 7-day window, download your output and then email us at oscar@optivox.app with the filename. We will delete it and confirm in writing, normally the same working day. You can also ask us to delete your account and everything attached to it at any time.

If your publisher or client requires it, we are happy to sign a mutual NDA or a confidentiality undertaking covering your work before you upload anything. We can also provide a written statement of the commitments on this page on Optivox letterhead, which is usually enough to satisfy a publisher's compliance check.

Email oscar@optivox.app and tell us what your agreement requires. Optivox is operated by Oscar Losseff, and you will be dealing with him directly.

If you are under an agreement that rules out third-party tools entirely, the straightforward way to evaluate Optivox is to process material you own outright: a sample chapter you recorded yourself, an out-of-contract title, or a few minutes of test narration in the same room, on the same microphone, at the same settings. That tells you what the enhancement will do to your actual recording chain without any protected file leaving your machine.

Files and account data are held with our hosting and processing providers, listed in our Privacy Policy, which may store data in the United States and European Union. Each provider handles data only as needed to deliver their part of the Service. Payment details never reach us. They are handled entirely by our payment processor.

If there is a specific clause in your agreement you need us to address, send it over and we will tell you plainly whether we can meet it. Email oscar@optivox.app.