AIoT edge computing solution
High performance, low power, low cost, small size, easy to use

AI Core Module
Product Features
- Small volume (25 * 25mm)
- Full pin out, up to 48 GPIO + 16 dedicated IO
- Optional onboard camera, onboard wifi (FCC / CE approved)
Development method
- The secondary development of C SDK, built-in common algorithm models (face recognition / posture recognition / flow statistics / ...)
- Serial port firmware, MaixPy firmware
Applicable scene
- Deep Custom Development End-Side AI Application
- Scenarios that require a lot of custom pin functions to facilitate secondary development
MF0 (AI Image Sensor)
Product Features
- Very small volume (20 * 20mm)
- 2.54mm DIP 8 + 6 pin / compatible SMD module, easy to debug and easy to integrate
- On-board camera mount, can be used by plugging directly into the camera
Development method
- Simple serial port firmware, 1KB memory microcontroller can also be used
- C SDK secondary development, providing common models (face recognition / posture recognition / flow statistics / ...)
Applicable scene
- Smart toys, fast AI-based upgrades of smart hardware
- Application of STEAM Education


MF1 (Living Face Recognition Module)
Product Features
- Can be adapted to a variety of screen sizes (1.3 / 2.4 / 2.8 / 3.5-inch screens), pin out, live face recognition module
- Onboard dual camera, IPS screen, Codec, RTC clock, speaker, relay, Wiegand interface, etc.
Development method
- Serial port firmware development, enhanced version of the serial communication protocol, both microcontroller and Android can be connected, serial port image transmission function
- Standard offline version firmware for easy access management
- Support SDK secondary development, packaged face recognition, network communication, voice broadcast, QR code scanning and other functions
Applicable scene
- Smart door lock, smart access control and other applications using small screen face recognition
MF5 (Large-screen living face recognition access control machine)
Product Features
- Can adapt to multiple screen sizes (4/5/7 inch screen), binocular live face recognition access control board
- On-board RTC clock, speakers, relays, Wiegand interface, etc.
Development method
- Standard offline firmware for easy access control management
- Standard WeChat firmware, easy management of WeChat applets
- Support SDK secondary development, packaged face recognition, network communication, voice broadcast, QR code scanning and other functions
Applicable scene
- Smart door lock, smart access control and other applications using large screen face recognition


MF5P
Product Features
- Can adapt to multiple screen sizes (4/5/7 inch screen), binocular live face recognition access control board
- On-board RTC clock, speakers, relays, Wiegand interface, etc.
- Onboard Linux coprocessor, users can customize the development GUI
Development method
- Provide linux BSP development kit, use QT to quickly develop your own GUI interface
Applicable scene
- High-performance large-screen face recognition application requiring complex human-computer interaction interface