DFRobot Launches HUSKYLENS 2: An Easy-to-Use AI Camera with Self-Training and Built-in MCP for LLM Connection

Click here to get this post in PDF

DFRobot today introduced HUSKYLENS 2, an intuitive AI vision sensor that empowers a global community of makers, educators, and AI enthusiasts to bring their ideas to life. This all-in-one device is built on a powerful 6 TOPS AI processor and is ready for immediate use with more than 20 pre-trained models for tasks like object tracking, hand recognition, and instance segmentation.

HUSKYLENS 2 is a pioneer, featuring a built-in Model Context Protocol (MCP) service that seamlessly bridges the gap between AI vision and large language models (LLMs). For projects requiring custom solutions, the sensor allows users to train and deploy their own AI models. With plug-and-play compatibility with popular platforms like Arduino, Raspberry Pi, and micro:bit, it is the ideal choice for robotics, automation, and STEM education.

Power Real-World AI with Robust On-Device Computing

At the core of HUSKYLENS 2 lies its exceptional computation power, featuring a dual-core 1.6GHz CPU, 6 TOPS of AI performance, and 1GB of memory. All algorithms run directly on-device, ensuring high-speed, low-latency results without reliance on the cloud. Whether using built-in models or user-trained algorithms, HUSKYLENS 2 delivers lightning-fast responsiveness and stable reliability. This enables real-time vision processing for projects in education, robotics, and automation—solving real-world challenges with ease.

Bridging Vision and Knowledge with MCP Service

HUSKYLENS 2 pioneers the integration of a built-in Model Context Protocol (MCP) service, revolutionizing how LLMs perceive the physical world. Through MCP, HUSKYLENS 2 seamlessly connects local visual recognition to large language models, allowing the camera to tap into the LLM’s extensive knowledge base. This elevates it from simply “seeing” objects or actions to truly “understanding” their context and implications. For instance, when glancing at your lunch, it uses MCP to leverage the LLM, transcending basic food identification to act as a nutritionist and provide personalized dietary advice.

Unlock Infinite Possibilities:Deploy Custom Models with YOLO

In addition to its built-in models, HUSKYLENS 2 supports the deployment of custom vision models trained using YOLO. The device is equipped with a comprehensive suite of annotation, training, and conversion tools, making it simple for users to create and deploy their own specialized models. This empowers makers to easily bring unique and personalized AI vision solutions into their projects.

Combine AI Models Flexibly for Smarter Projects

Breaking the limits of single-model execution, HUSKYLENS 2 enables flexible combinations and intelligent linking of multiple AI models. This opens endless possibilities for innovation—combine gesture recognition with object tracking, for example, to create hands-free, interactive projects.

Stream Real-Time Video Seamlessly via Wired or Wireless Connections

HUSKYLENS 2 features both wired (USB-C) and wireless (via an optional Wi-Fi module) real-time image transmission capabilities. Advanced image compression and stable protocols deliver low-latency, millisecond-level performance—allowing recognition results and video feeds to be accessed remotely in real time and powering intelligent devices with robust remote vision.

Adapt to Any Scenario with a Modular Camera Design

HUSKYLENS 2 features a modular camera system. By changing camera modules, it readily adapts to a wide range of scenarios, from macro object recognition to micro inspection or long-range monitoring. One device meets a full spectrum of visual sensing needs with unmatched flexibility.

Integrate Seamlessly Across Diverse Platforms and Systems

With standard UART and I2C protocols, HUSKYLENS 2 is truly plug-and-play across major platforms—including Arduino, micro:bit, Raspberry Pi, UNIHIKER and more. Its universal interface removes hardware limitations, giving users the freedom to build and innovate anywhere.

Key Specifications:

Processor: Kendryte K230 Dual-Core 1.6GHz

AI Computing Power: 6 TOPS

Memory: 1GB LPDDR4

Storage: 8GB eMMC

Image Sensor: GC2093, 2MP, 1/2.9″, 60FPS

Display: 2.4″ IPS Touch Screen (640×480), Full-Lamination

Other Components：1x Function Button, 2x LEDs, 1 x RGB LED

Wi-Fi: 2.4GHz Wi-Fi 6 (Slot-in Module)

Expandable Storage: 1 x TF Card Slot

Microphone: Capacitive Silicon Microphone

Speaker: 1W Speaker

Operating Voltage: 3.3–5V

Power Consumption: 1.5–3W

Dimensions:70 x 58 x 19mm

Interfaces: Type-C, 4-Pin Gravity, TF Card Slot

Driving Innovation with Purpose: Built for Real Needs

“At DFRobot, we created HUSKYLENS 2 for makers, educators, competition teams, and AI enthusiasts worldwide. Our mission with HUSKYLENS 2 is to make advanced AI vision more accessible than ever before: easy to use, yet incredibly powerful and adaptable for diverse applications. We believe HUSKYLENS 2 will empower users everywhere to turn their boldest ideas into reality, accelerate hands-on learning, and help the next generation lead innovation in AI.”

Available Now: A Valuable Addition at $74.9

Priced at just $74.9, HUSKYLENS 2 offers a professional-grade AI vision platform at an accessible cost. Interested users can purchase HUSKYLENS 2 now from the official DFRobot store.