What if tiny embedded devices could become intelligent AI companions with unique personalities?
That single idea turned into a small robotics project where I built two miniature AI robotic heads using Mini ESP32-S3 boards, OLED displays, speakers, and 3D printed parts.
The goal was not just to build another IoT device, but to create something that feels alive, interactive, and personal.
And honestly… seeing these tiny robots respond in realtime felt surprisingly futuristic.
Meet the Robots
I created two different AI personalities for this project.
🔵 ESP — Medical AI Assistant
ESP is designed as a small medical assistant robot capable of answering health-related questions and interacting like a calm, helpful AI companion.
Features:
- Medical-focused RAG knowledge
- Custom wake word: “Hi ESP”
- Realtime voice interaction
- Animated OLED face
- Compact robotic head design
🟠 Jason — Developer AI Assistant
Jason is built for developers and makers.
This assistant can help with coding discussions, embedded systems ideas, debugging concepts, and technical conversations.
Features:
- Developer-focused RAG memory
- Custom wake word: “Hi Jason”
- AI voice conversation
- Hacker/developer personality style
- Realtime communication
Why This Project Is Interesting
Most people think advanced AI assistants require powerful hardware or cloud servers.
But projects like this show something exciting:
Tiny embedded boards like the ESP32-S3 can already create surprisingly interactive AI experiences.
This combination of:
- Embedded systems
- Voice AI
- RAG memory
- OLED animations
- 3D printed robotics
- opens the door to affordable personal robotics.
Instead of building “smart devices,” we can start building:
- AI companions
- Learning robots
- Educational assistants
- Emotional interaction systems
- Personalized AI hardware
- That idea is what made this project so exciting for me.
Hardware Used
For this project, I used:
⚙️ Main Components
- Mini ESP32-S3 board
- OLED display
- Speaker module
- Wires & connectors
- 3D printed robotic head parts
- The ESP32-S3 works surprisingly well for lightweight AI interaction
systems because it provides:
- WiFi connectivity
- audio support
- compact size
- low power consumption
- enough processing for interactive applications
3D Printed Robot Heads
One of my favorite parts of the project was assembling the robotic heads.
The 3D printed enclosure gives the project a real personality instead of looking like just another exposed electronics prototype.
After assembling the:
- display
- speaker
- ESP32-S3 board
- wiring
the robot suddenly started to feel like an actual tiny AI companion.
Even simple OLED eyes and animations make a huge difference in giving life to embedded robotics projects.
AI + Wake Word + RAG
The most interesting technical part of this build is the AI interaction system.
Each robot uses:
✅ Different wake words
✅ Different AI personalities
✅ Different RAG memory systems
This means:
- ESP behaves like a medical assistant
- Jason behaves like a developer assistant
Even though both are running on similar hardware, their personalities feel completely different.
That’s the fascinating part of combining AI with robotics:
personality design becomes as important as hardware design.
Realtime Conversation Experience
Once everything was assembled and connected, I tested realtime conversation between the robots and users.
Watching a tiny ESP32-powered robot respond with voice output and animated expressions genuinely feels like something from the future.
Especially when:
- the wake word activates
- the face changes expression
- the speaker responds naturally
- the AI personality feels unique
It creates a much stronger emotional connection than a normal chatbot interface.
Full Build Video 🎥
I also documented the complete making process including:
- 3D printed assembly
- display wiring
- speaker integration
- ESP32-S3 setup
- Xiaozhi firmware
- AI assistant configuration
- realtime conversation demo
The video shows how everything was assembled step by step and how the final robots interact in realtime.
Top comments (0)