Keszoox
Voice Recognition Module V3 — 80 Commands, UART, Speaker-Dependent, Arduino Compatible
Voice Recognition Module V3 80 Commands, UART, Speaker-Dependent, Arduino Compatible
Couldn't load pickup availability
Voice Recognition Module V3 — 80 Commands, UART, Speaker-Dependent, Arduino Compatible
The Voice Recognition Module V3 is a compact speaker-dependent voice recognition board that stores up to 80 user-trained voice commands in onboard flash memory. Commands are trained by speaking into the onboard microphone and stored as acoustic templates. When a trained command is spoken, the module outputs the command index via UART to the host microcontroller (Arduino, ESP32, STM32). It is the most accessible voice control solution for Arduino-based smart home, robot control, and interactive electronics projects.
Key Specifications
| Parameter | Value |
|---|---|
| Recognition Type | Speaker-dependent (trained by user) |
| Max Voice Commands | 80 (stored in onboard flash) |
| Active Commands (simultaneous) | 7 (loaded into recognition group) |
| Interface | UART (9600 baud default) + SPI |
| Supply Voltage | 3.3V – 5V |
| Recognition Distance | ~1 meter (quiet environment) |
| Recognition Time | <100ms |
| Operating Temperature | -20°C to +70°C |
Voice Recognition Module V3 vs Google Assistant / Alexa
| Feature | VR Module V3 (this listing) | Google Assistant / Alexa |
|---|---|---|
| Internet required | No (fully offline) | Yes |
| Recognition type | Speaker-dependent (trained) | Speaker-independent (cloud AI) |
| Commands | Up to 80 custom commands | Unlimited (cloud processed) |
| Latency | <100ms | 500ms–2s (network dependent) |
| Privacy | Fully local, no data sent | Audio sent to cloud |
| Best for | Offline Arduino projects, custom commands | Smart home with internet |
Why Choose Voice Recognition Module V3?
- Fully offline — no internet, no cloud, no API key required
- 80 custom commands — train any word or phrase in any language
- UART interface — simple serial communication with any microcontroller
- <100ms recognition — fast response for real-time robot and device control
- 3.3V–5V supply — compatible with Arduino (5V) and ESP32/STM32 (3.3V)
Compatible With
- Arduino UNO/Nano: SoftwareSerial library for UART communication (pins D2/D3)
- ESP32: Hardware UART2 at 3.3V
- STM32: USART peripheral
- Raspberry Pi: UART serial port (/dev/ttyS0)
- VoiceRecognitionV3 Arduino library: Official library for training and recognition management
Frequently Asked Questions
Q: How do I train voice commands on the V3 module?
A: Use the VoiceRecognitionV3 Arduino library and the provided example sketches. (1) Upload the “train” sketch to Arduino. (2) Open Serial Monitor at 9600 baud. (3) Send the train command with the record index (e.g., “train 0”). (4) Speak your command clearly when prompted. (5) Repeat 3–5 times for better accuracy. (6) Load trained records into the active recognition group with the “load” command.
Q: How many commands can be active at the same time?
A: The V3 module can store 80 commands in flash but only recognizes 7 commands simultaneously (the active recognition group). To use more than 7 commands, implement a hierarchical command structure: use the first 7 commands as “menu” selectors, then load different command groups based on context. This allows effectively unlimited command trees.
Q: What is the recognition accuracy of the V3 module?
A: In a quiet environment at ~0.5m distance, recognition accuracy is typically 90–95% for well-trained commands. Accuracy decreases with background noise, distance (>1m), or if the speaker’s voice changes (illness, emotion). Train each command 3–5 times and avoid training in noisy environments for best accuracy.
Package Contents
- 1× Voice Recognition Module V3
- 1× Microphone (onboard or external, variant dependent)
