AI Disclosure
You sing. AI helps you get there. Every note that leaves this app comes from your voice — never a synthesised version of it.
At a glance
| Feature | Provider | What it does |
|---|---|---|
| Pitch detection | On-device DSP | Hears your singing in your browser. No audio leaves the device. |
| Voice profile | On-device analysis | Builds a numeric profile (range, key, stability) from your practice. Only numbers are stored. |
| Lyric suggestions | Google Gemini | Returns 5 alternative lines you can pick from, edit, or ignore. |
| Stage avatar | Google Gemini Image | Stylised stage poster from a selfie you opt-in to upload. |
| Coaching voice cues | ElevenLabs TTS | Reads a coaching prompt in a voice you choose. Not your voice. |
Your voice profile (Voice Lab)
When you train iPitch in Voice Lab, we analyse your singing in your browser and save a small JSON profile to your account containing:
- Lowest and highest comfortable note (Hz)
- Your home key (one of 12 pitch classes)
- Median pitch, stability score, sample count, timestamp
We do not store the raw audio, do not upload your microphone signal, and do not train any model on your voice. The profile is used to suggest a song key and melody range that fit you when you write in the Song Studio.
What we will never do
- We will never clone or synthesise your singing voice.
- We will never generate full songs sung in your voice.
- We will never train AI models on your microphone audio, recordings, or selfies.
- We will never sell your voice profile, lyrics, or images to third parties.
Why we don't offer voice cloning
Voice cloning means training an AI model on recordings of you so it can sing or speak any lyrics back in your voice. Other products offer this. iPitch deliberately does not, for four reasons:
- Deepfake risk. A cloned singing voice can be used to impersonate the owner — saying or singing things they never agreed to.
- Consent is hard to verify. We can't reliably prove the person uploading samples is the same person the voice belongs to.
- Store policy risk. Generative voice features get extra scrutiny in App Store and Play Store review and can block launches.
- It's not the product. iPitch helps you train your own voice. The win is you singing better, not a model singing for you.
If you want your voice on a backing track, our roadmap option is Backing Track Guide — we generate the instrumental in your key, you sing the vocal yourself.
Data retention & opt-out
- Voice profile: kept until you clear it from Voice Lab or delete your account.
- Lyrics & songs: kept in your account until you delete them.
- Selfies for avatars: stored privately. Removable from Account.
- Practice audio: never leaves your device. There is nothing for us to delete.
- You can export or permanently delete everything from Account settings at any time.
Age & consent
iPitch is intended for users aged 13 and older. If you are under the age of digital consent in your region, a parent or guardian must agree to these terms on your behalf. Voice Lab requires you to be signed in so a profile cannot be saved for anyone who has not consented.
Reporting AI-generated content
If lyric suggestions or generated avatars ever produce something that infringes your rights, is unsafe, or violates our Acceptable Use Policy, please report it to noetica@noeticaapps.com and we will investigate within 5 business days.
Not medical or therapeutic advice
iPitch is a practice and entertainment tool. It is not a substitute for a vocal coach, speech therapist, or medical professional. If you experience pain or strain while singing, stop and consult a qualified expert.
Related: Privacy · Terms · Acceptable Use · Contact