Optional reference photos
You may create a character using appearance choices without a reference photo. If you choose a photo, it can be resized and sent with the character request through our backend to a RunPod-hosted image-generation service. The image is used to guide a cartoon likeness. Local face/head-position detection also helps display generated characters. Photos and derived features are personal information; see the detailed privacy explanation.
Only use a photo you are authorized to provide. For another child, that requires appropriate parental authority; for an adult, obtain their permission. Do not upload sensitive or inappropriate images.
A grown-up’s familiar singing voice
An adult may record or import an audio sample and give permission to create a synthetic singing rendition. That rendition can sing words the person did not actually record. Saving an audio sample with a character can be local; selecting a voice for a connected song-generation flow can upload it for processing. The optional feature is available, as confirmed by the operator, but connectivity and native permission support vary by build.
The adult speaker must understand the purpose, processing and withdrawal options and agree to them. Do not use a celebrity, another person’s recording or a child’s voice without the necessary rights; child voice recreation is not supported. Choose a standard available singer if you do not want to provide an adult sample.
Processing and limits
Cloudflare/Supabase support private application records and media handling. RunPod-hosted services can receive the image/audio inputs needed for the generation stage. OpenAI handles personalized text/safety work and certain quality checks as explained in the Privacy Policy. These roles must be reflected in feature-specific in-app notice before the relevant disclosure.
Outputs can differ from the requested likeness or voice. Review them before child playback. Processing a recording for a familiar voice is not permission for public impersonation, advertising or unrestricted model training. The operator must verify the actual provider retention and independent-use restrictions before production use.
Removing an input or withdrawing permission
You can remove local photo/voice entries through the available profile/data controls. For backend records, provider copies or withdrawal of permission for further processing, email [email protected] with “Photo removal” or “Voice permission withdrawal”. Identify the parent account and saved item; do not attach the original recording or child photo to the request. An adult speaker whose voice was supplied by someone else may also contact us.
Removing a source does not necessarily remove an already-generated output, local archive or a copy exported outside the service. Tell us whether you also want related generated media removed. We will explain any remaining use or legally necessary record. See retention and data deletion.
Notice and consent are separate steps
This page explains the feature. It is not an agreement checkbox, recorded consent, identity check or substitute for a direct notice in the app. A production flow must give the person the relevant information and obtain any required separate consent before upload/processing. A general Terms acceptance does not authorize every optional use.