aidkit from neurocat

aidkit is a toolkit to test the perception functions of ADAS/ADS and ensure they are safe for deployment.

By testing the model at the heart of your perception component for robustness against a range of altered scenes, aidkit ensures reliable performance in your ODDs.

Where performance falters, aidkit guides your data campaign so you know how to improve.

aidkit further allows you to expand the range of situations your perception component can handle by combining the attributes of your ODDs to discover new safe and unsafe areas.

Built for ML engineers, aidkit integrates into existing workflows and is optimized to reduce time to deployment though useful features such as scaling, tracking, and automatic retraining.

Deploy a safer perception component with the empirical evidence to back up your safety claims to safety managers and external regulators.

For proven safer performance in more operational domains, aidkit ’til you make it.

built with your goal in mind

Getting your Perception System Safely on the Road

You need a tool to get your process moving so you can fulfill your safety requirements. 

aidkit was built with the same goal as you: safe deployment of your perceptions component in an ADS/ADAS.

Validating component safety with aidkit means reliable system performance in every-day and safety-critical scenarios.

This ensures you meet not only the exacting standards and expectations of your end-users but also your legal and regulatory obligations. 

aidkit is designed to make your tasks easier by integrating with your workflow and giving you the outputs that you need to meet your requirements.  

Contact us to see how we can help you build the AI systems of tomorrow and get it to market today. 

Your Requirements


Perception function safely enables tasks performance in all critical ODDs


System tasks can be performed under various scene conditions and corruptions


All critical ODDs, expected and unexpected, have been tested


Empirical evidence is available for safety arguments, validation, and audits


Enough data has been collected to enable testing and reporting

aidkit Capabilities


Efficiently test your perception component in numerous ODDs


Thoroughly test perception component against common and corner cases


Discover additional relevant ODDs by combining existing ODD attributes


Produce comparable and understandable statistical evidence


Guide an efficient data campaign that acquires and collects the right data



Software solution for safe perception components in automotive and mobility

symbolic image of solved puzzle


your perception component functions safely and reliably in every relevant ODD and scene regardless of what the environment throws at it

symbolic image of magnifying glass over document


safety with the documentation and empirical evidence needed to validate your claims to managers, regulators, and customers

symbolic image of rocket launching


without additional expenses and delays by ensuring your perception component is tested quickly but completely

aidkit Features

Extract risk scores for safety argumentation

Provide explainable and comparable evidence-based risk scores containing impact and probability of occurrence for every tested ODD scenario

Provide examples that challenge your perception component

Generate artifacts for a range of common and worst-case corruptions and adversarial attacks (i.e., perturbation-based image operations)

Discover and explore new ODD scenarios

Combine existent ODD attributes to automatically identify new safe and unsafe ODDs, as well as discover unexpected interaction effects and correlations

Tailor your data campaign and augment training & test data

Know exactly what data to collect and label to meet your performance shortfalls and leverage the corrupted images your model could not handle for re-training

Automate and accelerate perception component testing

Connect aidkit to your existing MLOps workflow via our Python-Client and scale your computing to execute tests faster

Track data lineage and fulfill audit requirements

Maintain an overview of existing models, model versions, data sets/splits and evaluations plus monitor interim pipeline results

neurocat Support

Bespoke solution projects

from aidkit integration to technical challenges of safety validation to results interpretation

Up-to-date tooling

reflecting the latest industry and safety standards even as they are written

Customized augmentations

from unique ODDs to tailored test images (corruptions) for your use case

A reliable partner

neurocat is riding shotgun with you on our journey from data to deployment

Want to learn more? Check out the aidkit website or get in touch with us!

neurocat logo