Hello,
I’m building a small learning app for children aged 4–9 (Korean literacy and early math). The first version is used only within my own family; a public release may come later.
I’d like to use gemini-3.8-flash-tts (paid tier) in the following way:
- I (an adult developer) generate audio files offline: instructions (e.g., “Try this addition”), hints, and the sounds of Korean syllables and words.
- The files are bundled with the app as static assets. The app never calls the Gemini API, and children never interact with Gemini in any way.
- No personal data about children is sent to the API. The input is only fixed instructional text that I write.
- As the father of two young children, I will personally listen to and review every generated clip that contains a sentence or a word before it is used. The only exceptions are clips that read a single number, syllable, or phoneme.
The Gemini API Additional Terms say the Services may not be used “as part of a website, application, or other service … that is directed towards or is likely to be accessed by individuals under the age of 18.”
My questions:
- Does the offline, adult-operated generation described above count as using the Services “as part of” a child-directed application?
- If it does, is there an approved path for child-directed educational apps — for example, the Gemini-TTS models offered through Google Cloud Text-to-Speech under the Google Cloud terms?
- Do the same age restrictions apply to Cloud Text-to-Speech voices that are not Gemini-based (e.g., Chirp 3 HD)?
I found similar questions on this forum that are still unanswered:
An official clarification would help other developers of educational apps as well.
Thank you.