Discover NanoEdge AI Studio, the market leader for EDGE AI

A search engine desktop software for AI libraries, designed for embedded developers.

With Nanoedge AI Studio, find the best AI library for your embedded project, and start incorporating machine learning capabilities into the C code in your MCU.

Documentation, tutorials, videos, FAQ, and much more...

Features
NanoEdge Studio

Windows 10 and
Ubuntu versions

Your production data is never sent to a cloud

Dedicated to
embedded developers

Libraries can run on every
Arm© Cortex-M microcontroller

Automatic data
quality check

Automates the search for the best AI models for your project

Collect and import data in real time via serial port

Emulator to test a library before embedding in a connected object

C libraries that are easy to deploy

NanoEdge AI Studio simplifies Machine Learning and Signal Processing 

Input

Project parameters + signal examples for context

Output

Precompiled static library (.a) to link to your main code

Discover our
software
in 5 steps

DISCOVER
discover
DISCOVER
DISCOVER
DISCOVER
Previous
Next

NanoEdge AI Library

Our machine learning models are developed in-house, from the ground up, starting with the algebra. Several models are built from scratch, while others are based on traditional AI/ML models (e.g. kNN, SVM, neural networks…).

Each Nanoedge AI Library is optimised thanks to several user-imported signal examples. It contains the best ML model for your project, combined with adequate signal pre-processing and optimal hyperparametrization. This library provides simple and intuitive functions (learn, detect, classify) that bring powerful ML features to any Cortex-M C code.

Signal pre-processing

ML model

Hyperparameters

AI library

Designed
for microcontrollers, our models:

Are optimised to run on MCUs (every ARM Cortex-M)

Are very memory efficient (1-20KB RAM/Flash)

Are ultra fast (1-20ms inference on M4 80MHz)

Can be trained and used directly within the MCU

Can be integrated into
existing code / hardware

Consume very
little energy

Preserve your stack (static allocation, no dynamic allocation)

Do not rely
on the cloud

Do not require any
ML expertise for creation and deployment

2 types of librairies
available

Anomaly detection library

Based on « in situ » training (directly inside Cortex-M MCU). This kind of ML library can be used to learn a series of nominal signals and detect anomalies based on that model.

Classification library

Based on « in studio » training. This kind of library can classify different classes of signals inside a Cortex-M MCU using ML.