The AI world is rapidly evolving, but let’s face it – traditional AI model training and deployment can be overwhelming, especially for beginners who are targeting to use the models for their AI applications. While there are amazing platforms out there trying to simplify things, many still require technical expertise.

Inspired by Google Teachable Machine and YOLO World, SenseCraft AI is a free, web-based platform designed for beginners, focuing on a no-code approach and an application-orientation. Our goal is to simplify and accelerate the creation of AI applications, making the process beginner-friendly and accessible to all.

What Makes SenseCraft AI Special?

Feature 1: Built For Beginner

We’ve made everything web-based. No troublesome installation or configuration is needed. You can visit the SenseCraft AI platform from the Chrome or Edge browser anywhere in the world.

Understanding the starting points of most of the AI beginners, SenseCraft AI also offers an extensive model zoo with over 400 pre-trained models ready for immediate deployment. This resource is perfect for users needing quick solutions without the time for training.

Feature 2: Truely No-Code

The no-code approach works for both model training and deployment. For deployment, you’ll only need a few clicks to get your model running on your device.

For model training, we support two types of models without any coding.

Classification Models (2 simple steps):

• Collect and Classify Your Scenes: Hold to capture around 10 photos for each class, and then group your examples into distinct classes (like “room occupied with lights on” vs. “empty room with lights off”)
• Train with One Click: Hit train and instantly preview your model’s accuracy

Object Detection Models:

Our exciting beta feature lets you create detection models by simply typing in the object name! While we’re still improving the accuracy, you can enhance results by adding your own training images.

Feature 3: Application Focused

We’ve gone beyond just model training and deployment. SenseCraft AI supports multiple output methods (MQTT, GPIO, UART, I2C, SPI) to stream your AI inference results with any microcontrollers or systems. For instance, once you’ve deployed a classification model onto Grove Vision AI Module V2, and then you can port out the AI classification results via UART to an Arduino Uno, and even triger a specific event (eg. activate another connected relay to open the door for a detected classification). For a more detailed live demo, I joined the livestream on YouTube with Violet to build a LVDuo Recognizer last week. You can see the demo as follow.

Moreover, the Vision Workspace on SenseCraft AI also guides you through every step, from model deployment to setting up outputs, ensuring a beginner-friendly experience.

Hardware You Can Use

Currently, six Seeed product lines work seamlessly with SenseCraft AI, ranging from microcontrollers, AI gadgets, IP66 rating sensor devices to powerful edge computing devices. They are XIAO ESP32S3 Sense, Grove Vision AI Module V2, SenseCAP Watcher, reCamera, SenseCAP A1102 LoRaWAN Vision AI Sensor, and reComputer Jetson Edge Computing Devices.

But that’s just the beginning! We’re adding more hardware support and enabling model downloads in various formats for compatibility with other brands.

Embrace Your AI Journey Now

We’re excited about making AI accessible to everyone, and we’d love to have you join us on this journey. Share your ideas for new models, suggest AI frameworks, or tell us about your experience through our GitHub repository or wiki.

What do you think about SenseCraft AI? Drop your thoughts or questions below!