Skip to main content

Worldwide Shipping · Secure Payments

AI Summary

The Coral Dev Board Micro is a low-power AI microcontroller with a dual-core Cortex-M7/M4, Edge TPU, camera, and microphone. It enables fast on-device machine learning for vision and audio projects.

Key Highlights

  • Dual-core Cortex-M7/M4 MCU with Coral Edge TPU for efficient ML
  • 4 TOPS int8 performance at 2 TOPS per watt
  • 64MB RAM for model and application data
  • Integrated Himax color camera and PDM mono microphone
  • Two 12-pin GPIO headers and two 100-pin board-to-board connectors
  • USB Type-C (USB 2.0) for power and data

Ideal For

  • Embedded ML developers prototyping low-power vision systems
  • IoT engineers building smart sensor nodes with local inference
  • Makers and hobbyists exploring TensorFlow Lite on microcontrollers
  • Researchers testing audio and vision models at the edge
  • Product designers creating compact AI-enabled devices

Buying Tip

Consider whether you need the Wi-Fi/Bluetooth or Ethernet/POE add-on board for your connectivity requirements, as neither is included by default.

Elecvora expert review

Our verdict

The Coral Dev Board Micro is a well-integrated platform for edge ML prototyping, combining a capable MCU, Edge TPU, and sensors in a compact form. It is best suited for developers who need efficient on-device inference and are comfortable with embedded workflows.

AI-assisted analysis based on the manufacturer's specifications. Not a hands-on lab test.

Pros

  • Integrated Edge TPU delivers fast, efficient inference
  • On-board camera and microphone simplify vision and audio projects
  • Dual-core MCU enables cascaded low-power processing
  • Expandable with board-to-board connectors

Cons

  • Limited RAM (64MB) may constrain larger models
  • No wireless connectivity built in; requires add-on board
  • Requires separate USB-C cable and host for development

Best for

Embedded ML developers and IoT engineers

Not ideal for

Beginners seeking a plug-and-play experience

Specifications

Accelerometer TypeCoral Edge TPU coprocessor: 4 TOPS (int8); 2 TOPS per watt
RAM64 MB
SensorHimax color camera (324 x 324 px); PDM mono microphone
Input Output2x) 12-pin GPIO header; (2x) 100-pin board-to-board connectors; USB Type-C (USB 2.0)
User Interface(4x) LEDs; (2x) switch buttons
Accessory TypeCoral Wi-Fi/Bluetooth Module Board; Coral Ethernet/POE Board
Shipping Weight0.028 kg
Shipping Dimensions14 × 11 × 1 cm

Frequently Asked Questions

What is the Coral Dev Board Micro used for?
It is used for prototyping and deploying low-power embedded machine learning applications, such as vision and audio inference, directly on the device.
Does the board include Wi-Fi or Bluetooth?
No, the base board does not include wireless connectivity. You can add the Coral Wi-Fi/Bluetooth Module Board via the board-to-board connectors.
What power input does the board require?
The board is powered via its USB Type-C port (USB 2.0). Specific voltage requirements are not listed in the product data; use a standard USB power source.
What is included in the box?
The package contents are not specified by the manufacturer. Typically, only the board is included; please verify with the seller.
Can I connect external sensors and actuators?
Yes, the two 12-pin GPIO headers allow you to connect external peripherals. Additionally, the 100-pin board-to-board connectors support add-on boards.
How should I care for the board?
Store it in a dry, static-free environment and handle it by the edges to avoid damage. Use a proper USB-C cable and avoid exposing it to extreme temperatures or moisture.

Overview

The Coral Dev Board Micro is a microcontroller board built for embedded machine learning at the edge. It pairs a dual-core Cortex-M7/M4 processor with the Coral Edge TPU, a dedicated accelerator for TensorFlow Lite models. This combination allows you to run vision and audio inference directly on the device, reducing latency and preserving data privacy. The board includes a Himax color camera (324 x 324 px) and a PDM mono microphone, making it ready for sensing applications right out of the box.

With 64MB of RAM and a compact form factor, the Dev Board Micro is suited to low-power systems where space and energy efficiency matter. It provides two 12-pin GPIO headers and two 100-pin board-to-board connectors for expanding hardware with custom add-on boards, such as the Coral Wi-Fi/Bluetooth Module Board or the Coral Ethernet/POE Board.

Key Features

  • Edge TPU coprocessor: Delivers 4 TOPS (int8) and 2 TOPS per watt, enabling fast, power-efficient inference.
  • Dual-core MCU: Cortex-M7 and Cortex-M4 cores handle control tasks and can cascade from ultra-low-power to more complex processing.
  • On-board sensors: Himax color camera and PDM mono microphone support vision and audio ML applications.
  • Expansion options: Two 12-pin GPIO headers and two 100-pin board-to-board connectors for add-on boards.
  • User interface: Four LEDs and two switch buttons for status and input.
  • Connectivity: USB Type-C (USB 2.0) for power and data transfer.

Specifications Explained

The Edge TPU is the key component for machine learning. Its 4 TOPS performance means it can execute four trillion int8 operations per second, which is sufficient for many real-time vision and audio models. The 2 TOPS per watt figure highlights its efficiency, making it suitable for battery-powered or energy-harvesting designs.

The 64MB RAM provides memory for model storage and runtime data. The dual-core Cortex-M7/M4 architecture allows the M7 to handle heavier processing while the M4 manages low-power tasks, enabling a cascaded approach where simple inferences run on the M4 and more complex ones on the M7 or Edge TPU.

The Himax camera outputs 324 x 324 pixel color images, adequate for tasks like presence detection, object counting, or simple classification. The PDM mono microphone captures audio for keyword spotting or sound event detection. The two 12-pin GPIO headers expose pins for external sensors and actuators, while the 100-pin board-to-board connectors allow stacking add-on boards for wireless or wired networking.

Compatibility and Requirements

The board supports TensorFlow Lite models and is designed to work with the Coral ecosystem. It can be expanded with the Coral Wi-Fi/Bluetooth Module Board and the Coral Ethernet/POE Board via the board-to-board connectors. A USB Type-C cable is required for power and data connection to a host computer. No other compatibility details are specified.

Typical Uses

  • Prototyping low-power vision systems such as smart cameras or presence detectors.
  • Audio event detection and keyword spotting using the on-board microphone.
  • Embedded ML projects that require local inference without cloud dependency.
  • Educational and research projects exploring edge AI and TinyML.
  • Custom IoT devices with add-on networking boards.

What's in the Box

Package contents are not specified by the manufacturer. Please check the product packaging or contact support for details.

Quick specs

Accelerometer Type
Coral Edge TPU coprocessor: 4 TOPS (int8); 2 TOPS per watt
RAM
64 MB
Sensor
Himax color camera (324 x 324 px); PDM mono microphone
Input Output
2x) 12-pin GPIO header; (2x) 100-pin board-to-board connectors; USB Type-C (USB 2.0)
User Interface
(4x) LEDs; (2x) switch buttons
Accessory Type
Coral Wi-Fi/Bluetooth Module Board; Coral Ethernet/POE Board
See all specifications