M5Stack ASR Module (CI1302)

The M5Stack Module ASR is an offline CI1302 voice module with microphone, speaker, echo cancellation and mid-session interruption. It includes 53 English commands preloaded, supports UART or keyword activation, and can be customised up to 300 commands across four languages.

---

The M5Stack Module ASR is an offline AI voice recognition module based on the CI1302 chip, designed for responsive voice control without needing an internet connection. It includes an onboard microphone for clear audio capture and a built-in speaker for high-quality audio feedback, making it suitable for voice-driven interfaces and smart device control.

To improve usability in real-world environments, the module supports mid-session voice interruption, allowing you to interrupt recognition and issue a new command without waiting for the previous session to finish. It also includes AEC (Echo Cancellation) to reduce echo and background noise interference, improving recognition accuracy.

The module ships with 53 English wake-up and feedback command words preloaded and can be activated either by UART commands or by voice keyword triggers. For more advanced projects, users can regenerate the firmware to customise recognition words in Chinese, English, Japanese, and Korean, supporting up to 300 command words.

Communication is handled via UART, and a built-in DIP switch allows switching between different UART pin configurations, making integration more flexible across different boards and setups.

Module ASR is well suited to applications such as AI assistants, smart homes, security monitoring, automotive systems, robotics and smart hardware, and healthcare devices, offering a practical solution for offline voice interaction.

Features

• Intelligent offline voice module
• AEC noise reduction processing
• Built-in 53 English commands; support up to 300 commands
• UART communication interface
• Support for custom wake-up words
• Firmware burning via reserved USB Type-C interface
• Equipped with microphone and built-in speaker
• Development Platform
  • UiFlow2
  • Arduino

Specifications

AI Voice Chip	CI1302
Command Words	Up to 300 commands supported; 53 preset commands from factory
Wake-up Method	Voice keyword or UART serial communication to activate device
Microphone	Analogue microphone, model GMI4527P-2C-32db
Speaker	Chamber speaker 8Ω@0.8W (used for wake-up feedback output)
Communication	UART serial communication, default baud rate: 115200@8N1
Noise Reduction	AEC echo cancellation noise reduction
Wake-up Range	Ambient noise 40dB: up to 6.4m Ambient noise 54dB: up to 1.8m
Power Consumption	Standby: DC [email protected] Working: Low volume: DC [email protected] Medium volume: DC [email protected] High volume: DC [email protected]
Product Size	54.0 x 54.0 x 13.1mm
Product Weight	15.2g

Resources

• Documents

Package Contents

• 1x ASR Module