The robot is mounted on a tilting platform that can incline any device up to a 30-degree angle. This can be used to switch the device from landscape to portrait mode, a capability which allows for testing complex applications like the ones used for AR and VR.


MATT can tackle any tests that require mandatory physical input. It can press physical buttons, side buttons and any other data input form. The testing robot will be an excellent tool for any product where the input cannot be easily simulated, like PoS or systems where the same input must be introduced multiple times.


Many of the interactions with a touchscreen can be simulated with software testing platforms like Selenium, but due to the wide variety of hardware devices, this can be a complicated task. However, MATT manages both HW and SW interaction with ease. Due to MATT’s flexibility, it can be integrated into any testing framework, therefore synchronization between software testing automation and hardware button pressing can be achieved seamlessly.

MATT effector


MATT can be equipped with an external high-speed camera for different measurements. Moreover, the speed of up to 2m/s ensures that the robot can reach any point swiftly and with high accuracy.


Building on high accuracy, MATT moves with 0.05 precision and touches with exact pressure, capabilities that make the robot an ideal tool, effective in touch screen device calibration and not only. To better fit use cases, the capacitance, size and shape of MATT fingers can be customized.


The robot can press any buttons, placed vertically or sideways, and interact with any touch screen devices (resistive or capacitive).


MATT is a flexible asset, whose potential can be directed towards the specific needs of a company and use case.

Used as a tool or as a complete HW & SW package

MATT’s design is focused on modularity; for this reason, the hardware platform and the API can be seen as different parts. Depending on the application, MATT can be employed either as an embedded tool, performing basic, yet powerful tasks, or as a complete, unified solution.

Integration with A.I. /Machine Learning Frameworks

MATT’s cross-platform software architecture allows it to easily integrate with frameworks dedicated to Machine Learning for image processing. One successful example is Intel’s OpenVINO toolkit, which has enabled MATT to solve captcha tasks, such as selecting the cars from a collection of distinct images. When it comes to integration with AI or Machine Learning Frameworks, MATT’s applications are unlimited.

The most complete solution for physical devices testing

From detecting production faults and performing calibration processes, to identifying and reporting software inconsistencies, as well as executing more intelligent & challenging tasks, MATT offers all the functionalities underlying physical device testing.

Device OS Independent

When developing MATT, our philosophy was to use computer vision in order to solve the most difficult tasks. Because of that, our solution supports all operating systems.
MATT OS independent devices

Open Ecosystem

Advanced users can reference the extensive, low-level serial protocol documentation to obtain maximum control over the system’s movements and trajectories. The embedded, industrial USB3 Vision camera is accessible to third-party computer vision toolkits like NI Vision Builder™, OMRON AutoVision™ and many more.

Remote - Controlled

Every MATT robot can be controlled independently, on the network, using the included software suite. Its powerful and easy-to-use Python API offers adaptability in controlling the testing process in both multi-robot and multi-device scenarios simultaneously, achieving unmatched flexibility.

Maximum Speed (M/S)
Precision (MM)
Degrees of Freedom

“MATT’s idea was born as soon as we faced the challenge of testing multiple devices and their constant software updates quickly and accurately. None of the available tools were fully capable of satisfying our demand for efficiency, which is the same demand of many other companies. We decided to build it ourselves, starting from our in-depth knowledge of the processes and requirements, creating a machine that would qualify in terms of reliability, accuracy and scalability.”