KidQA

About this project

KidQA is a voice-first Q&A app for children ages 3 to 6. Tap the mic, ask a question out loud, and get a short, age-appropriate spoken answer with a matching picture. No reading or typing required. Behind the scenes, browser speech recognition transcribes the question, a language model writes the answer, and streamed text-to-speech reads it back while the image loads in parallel.

A realtime mode goes further: audio streams directly between the browser and the OpenAI Realtime API over WebRTC, so the model listens, thinks, and speaks over one persistent connection, which makes the conversation feel noticeably faster and more natural. Separate chained modes swap in different answer models, so providers can be compared side by side on identical questions.