Meet Pico.
Your AI companion that sees, hears, and feels.
An emotionally responsive desktop robot that communicates like a pet — through chirps, expressions, and movement.
What Is PICO?
More than a smart speaker. A companion that truly responds.
“Unlike smart speakers that just answer questions, Pico behaves like a living pet. It's a non-verbal AI companion that communicates through expressive sounds, animated eyes, and head movements.”
It Sees You
Face detection and recognition. Knows your face, remembers you.
It Hears You
Wake-word detection and speech-to-text. Understands your commands.
It Feels Touch
Capacitive touch sensor. Pet it and it purrs.
Default state
Subtle breathing sounds
A Personality That Breathes
PICO's Emotion Engine is a state machine that processes inputs and transitions between emotions — just like a living creature.
Emotion States
How It Works
- →Sensory inputs (camera, mic, touch) trigger emotion transitions
- →Each state has its own eye expression, sounds, and head movement
- →Transitions are smooth — half-blink between expressions
- →Idle behaviors like random blinking and pupil drift add life
What Pico Can Do
A complete AI system packed into a tiny companion.
Wake Word Detection
"Hey Pico" trigger via ESP-SR, with offline fallback.
Speech-to-Text
Google Cloud Speech-to-Text with 60-min free tier.
Natural Language Processing
Google Gemini for conversational AI and contextual responses.
Sound Bank Communication
20+ expressive WAV files — no TTS, pure personality-driven audio.
Software-First Approach
Perfect your AI on PC before spending a single rupee on hardware.
Simulate on PC
Develop and test 100% of AI functionality on your laptop using Python. Camera, microphone, and speakers simulate the robot hardware. Zero hardware cost to start.
- Python 3.11 + OpenCV + face_recognition
- Google Cloud Speech-to-Text + Gemini AI
- Sound Bank with 20+ expressive WAV files
- Full Emotion Engine state machine
Port to Hardware
Translate validated Python logic to C++ (Arduino framework). Upload firmware to the ESP32-S3-EYE via USB. Same algorithms, embedded performance.
- Arduino IDE + ESP-IDF toolchain
- ESP-WHO face detection on-device
- ESP-SR wake-word recognition
- I2S audio + OLED display drivers
Integrate & Build
Assemble the physical robot with 3D-printed shell, servo head tracking, touch sensor, and speaker. Total hardware cost under ₹10,000.
- ESP32-S3-EYE + 0.96" OLED display
- Pan-tilt servo head (2× SG90)
- MAX98357A I2S amp + speaker
- 3D-printed magnetic shell
Built on ESP32-S3
Everything you need, at an accessible price point.
Total estimated cost • Indian market pricing • Verified suppliers
ESP32-S3-EYE
- Dual-core 240MHz processor
- 8MB PSRAM + 16MB Flash
- Built-in 2MP camera + microphone
₹4,200 – ₹5,500
0.96″ OLED (SSD1306)
- 128×64 I2C display
- Self-emissive — no backlight
- Perfect for expressive eye animation
₹150 – ₹300
2× SG90 Micro Servo
- 180° rotation range
- 1.8 kg·cm torque
- Pan-tilt head tracking assembly
₹200 – ₹400
MAX98357A Amplifier
- I2S digital audio interface
- 3W output power
- No DAC required — direct ESP32 connection
₹180 – ₹350
3W Speaker (40mm)
- 4Ω impedance
- Full-range driver
- Clear sound for chirps and beeps
₹50 – ₹150
Touch Sensor (TTP223)
- Capacitive touch detection
- Single-pin digital output
- Pet-petting interaction trigger
₹30 – ₹60
The Tech Inside
A complete AI stack — from Python simulation to embedded firmware.
PC Simulation
Phase 1 — Python 3.11
- OpenCV— Camera & image processing
- face-recognition— Face detection & identification
- sounddevice— Audio recording & playback
- Google Cloud STT— Speech recognition (free tier)
- Google Gemini— Conversational AI
Cloud Services
Free Tier APIs
- Speech-to-Text— 60 min/month free
- Gemini AI— 60 req/minute free
- WiFi Bridge— ESP32 ↔ Cloud
Robot Firmware
Phase 2 — C++ (Arduino)
- ESP-WHO— On-device face detection
- ESP-SR— Wake-word recognition
- FreeRTOS— Real-time OS multitasking
- I2S Audio— Digital audio output
- SSD1306 Driver— OLED display rendering
Ready to Build?
Start with zero hardware cost. Perfect your AI on your laptop, then bring it to life on the ESP32.