When OpenAI is used
In production, Wake Up Drawing uses OpenAI for story text, safety checks, and optional background images, but only after an active grown-up step. The family owner can disable story generation, and every request requires PIN approval.
Text requests contain the grown-up-confirmed fictional label, allowlisted character traits, structured story choices, language, length, and prose level. Generated text may be sent back for safety, quality, and bounded repair checks. In Classic, image requests contain a structured background scene without the main character; the generated background is then checked for safety and quality.
A raw camera photograph, its surroundings, real name, birthdate, school, address, contact details, and other private child information must not be sent to OpenAI. The app asks the grown-up to confirm this before generation.
Classic sends no drawing to the image model. Creative requires a fresh notice and exact approval of the cleaned metadata-free sprite. Immersive separately requires approval of the exact metadata-free whole drawing and is limited to six pages. Both are optional, separately labelled, owner-enabled for selected profiles, and require grown-up review of the result.
Text requests use store: false. OpenAI states that API data is not used for model training by default unless the customer opts in to data sharing, but default abuse-monitoring logs may retain customer content for up to 30 days. Wake Up Drawing does not claim Zero Data Retention. See the OpenAI API data controls.
For image endpoints, inputs and outputs may appear in abuse-monitoring logs for up to 30 days under the selected standard posture. OpenAI also documents exceptional longer retention of image inputs flagged for a possible policy violation. The image endpoints otherwise keep no application state for this processing.