AI Robot Toy Development
Took a consumer AI robot toy from concept to full-scale manufacturing-ready product — combining embedded hardware design, real-time audio/video streaming, and a custom AI backend into a seamless conversational experience.
Project Overview
Took a consumer AI robot toy from concept to full-scale manufacturing-ready product — combining embedded hardware design, real-time audio/video streaming, and a custom AI backend into a seamless conversational experience for end users.
Built on an ESP32-S3, the robot lets users speak naturally — microphone audio is live-streamed to a cloud AI server via WebSocket, processed, and the AI-generated response (audio + animated visuals) is streamed back in real-time chunks and played out through the integrated speaker and LCD screen simultaneously. Chunked streaming was essential to minimise perceived latency on a resource-constrained embedded platform, keeping the conversation feel natural and responsive. Acoustic Echo Cancellation (AEC) was implemented in the microphone circuitry to prevent speaker output from feeding back into the mic — a critical hardware-software challenge for always-listening voice devices.
A custom BMS and battery management circuit powers the robot safely for extended play sessions. WiFi handles the primary AI streaming pipeline while BLE enables seamless pairing with a companion mobile application for setup, parental controls, and personalisation. The full system — hardware PCB, firmware, streaming protocol, and mobile app integration — was delivered manufacturing-ready.
Tech Stack
Gallery
Gallery images coming soon.