The Intelligence Inside
A look under the hood at the custom-built AI and privacy-first architecture that makes In Already the most advanced smart kitchen system on the planet.
Our Three-Model AI Core
The Grocery Detector
Our foundational object detection model, fine-tuned on a massive, curated dataset of over 200,000 grocery and kitchen items. It's the system's eyes, capable of identifying hundreds of distinct products in a cluttered, real-world fridge environment.
The Produce Classifier
A specialized classification model trained exclusively on fresh produce. This expert model can distinguish between different varieties of apples or types of lettuce, providing a level of detail that general models miss.
The Date-Block Finder
A highly-focused detection model trained on the ICDAR text dataset. Its sole purpose is to locate the exact position of expiry date blocks on any package, which it then passes to our OCR engine for precise, accurate reading.
Privacy by Design
Unlike other smart devices, your data is yours. The In Already camera processes everything on-device. The powerful onboard chip runs our custom AI models locally. No images of your kitchen are ever uploaded to the cloud. The only data that leaves the device is anonymous and statistical (e.g., "1 apple identified") to help us improve the global model for everyone.
A Self-Hosted Culinary Brain
Our recipe suggestions aren't pulled from a simple online search. We built our own self-hosted recipe engine from the ground up, using the RecipeNLG dataset with over 2 million recipes. This allows our AI to perform complex, multi-ingredient searches instantly, finding creative and delicious meals that other services would miss, all without relying on external APIs.
Built on the Best
We believe in the power of open-source. Our entire software stack is built on powerful, community-driven technologies like Python, YOLOv8, and PyTorch, allowing us to innovate faster and pass those benefits on to you.