Evidence at a glance
| Question | Current evidence |
|---|---|
| Voice feature | Marketed by OhChat as part of creator experience |
| Quality score | Pending live account test |
| Generation delay | Pending live test — record time from request to delivery |
| Creator likeness | Pending — compare with known public voice sample |
| Allowance / cost | Verify in logged-in account before publishing |
| Privacy | Voice requests are processed by OhChat systems |
Voice test protocol
Use this five-point SFW protocol for an honest evaluation:
- Greeting: Request a casual greeting — test naturalness and pacing
- Emotion: Request a warm or excited response — test emotional range
- Pronunciation: Include a proper noun or unusual word — test accuracy
- Repetition: Request the same phrase twice — test consistency
- Delay: Time from request to audio delivery — record in seconds
What to measure
- Generation time (request → audio file available)
- Audio quality (clarity, artefacts, background noise)
- Pronunciation accuracy (especially names and specific words)
- Creator likeness (for Digital Twins — compare with approved public sample)
- Messages per allowance and cost per extra message
Privacy considerations
Voice requests are processed by OhChat's systems. For creator Digital Twins, voice may use creator-approved cloning technology. Treat voice interactions as cloud data under the Privacy Policy. Do not record or redistribute voice outputs without reviewing the Terms.
⚠ Publication gate: Do not publish a quality score until live tests are complete with dated evidence. Do not republish voice clips of real creators without permission.
Related guides
Frequently asked questions
Voice is marketed by OhChat. Quality, allowance and cost require account-level verification.
For Digital Twins, voice may use creator-approved cloning. It is an AI voice, not the human speaking live.