HUSKYLENS 2 is a next-generation AI vision sensor built to make advanced machine vision accessible to every creator. Powered by a 1.6 GHz dual-core processor and a 6 TOPS on-device AI accelerator, it runs all algorithms locally for fast, low-latency performance. Features Fast On-Device AI At the heart of HUSKYLENS 2 is a dual-core 1.6 GHz processor paired with a built-in AI acceleration unit capable of 6 TOPS. Combined with 1 GB of RAM, this robust hardware allows the device to execute complex algorithms and neural networks entirely on-device. As a result, HUSKYLENS 2 can deliver real-time performance for intensive tasks such as multi-object detection and detailed pose estimation, achieving high speed and low latency without relying on cloud processing. 20 Built-in AI Models HUSKYLENS 2 is built for instant productivity. It arrives pre-loaded with over 20 AI models that support a broad range of common computer vision tasks, from face recognition to object tracking. With this robust library ready to go, users can begin creating projects immediately and with minimal setup, making AI application development more accessible than ever. Empowering LLMs with Contextual Awareness via MCP Service HUSKYLENS 2 is the first AI vision sensor to integrate a built-in MCP service, redefining how LLMs interpret the physical world. Instead of supplying raw image snapshots like traditional cameras, HUSKYLENS 2 provides structured, intelligent summaries. It can tell a Large Language Model (LLM) that it is observing a specific, pre-identified individual performing a particular action. This gives LLMs precise, context-aware perception, enabling them to make smarter and more informed decisions. Deployment of Custom Models Trained with YOLO Users can also create their own vision models with YOLO and deploy them on HUSKYLENS 2. The device includes a full suite of annotation, training, and model-conversion tools, making it simple to build and deploy custom models. Flexible Model Combination for Diverse Applications The HUSKYLENS 2 architecture supports running multiple models either in sequence or simultaneously, enabling the development of more advanced and nuanced applications. For example, hand-keypoint detection can be paired with object tracking to build sophisticated contactless control systems, where a gesture triggers the tracking of a chosen object. This multi-model capability makes it possible to design layered logic and more intelligent interactive behaviors. Real-time Video Streaming The HUSKYLENS 2 enables effortless data collection and image viewing through real-time video streaming. It supports image transmission via both wired USB-C and an optional wireless Wi-Fi module. This capability is essential for remote monitoring, building datasets for training new models, and debugging visual applications by offering a live view of the sensor’s output. Replaceable Camera Module HUSKYLENS 2 features a modular camera system designed for effortless adaptation across diverse applications—from microscopic inspection to long-distance monitoring. Its highly flexible design supports a variety of interchangeable lenses, including manual-focus, microscope, and night-vision modules. Broad Compatibility HUSKYLENS 2 uses standard UART and I2C interfaces, making it easy to connect with many microcontrollers and single-board computers, including Arduino, micro:bit, Raspberry Pi, ESP32, UNIHIKER K10, and UNIHIKER M10. Contents 1x HUSKYLENS 2 AI Vision Sensor 6x M3 Screws 6x M3 Nuts 1x Mounting Bracket 1x Heightening Bracket 1x Gravity-4P Sensor Connector Cable (30cm) 1x Dual-Plug PH2.0-4P Silicone Cable (20cm) 1x Power Adapter Board
