Deploy Locally and Centrally. Execute Everywhere.

Distributed Real-Time Analysis and Equipment Monitoring
Stream Analyze specialize in software for scalable real-time analyses of data in motion.

Our analytical software, sa.engine, can be deployed on to any hardware device, both locally on the edge or centrally in an arbitrary cloud solution.

Stream analyze provides a simple and straight-forward tool for quickly and easily deploying ad-hoc mathematical, statistical and AI models to edge devices. The solution can be deployed stand-alone or complementing current solutions, providing them with unique functionality. Due to our small footprint (<1 MB), we can deploy on nearly any device.

sa.engine can also be deployed centrally to an arbitrary cloud solution, providing for orchestrated device analysis and developing mathematical and AI models. Our central solution can also be used for providing centralized device monitoring, as used in monitoring centers. Ad-hoc analysis allows staff to follow up on any alarms and notifications from devices, drilling down on what is actually happinging in the field.

Our mobile analysis allows sa.engine to be deployed to any laptop, handheld or mobile device. Thus, technicians can perform specific ad-hoc anlysis prior to visiting a facility or site.

For the Analyst Everywhere

Empower analysts with a high-level on-line analysis and query language where expressions and algorithms can immediately be executed. There is no need for programming skills whatsoever.

The flexibility of the our tool provides the possibility to execute ad hoc analysis across all environments simultaneously, locally or centrally.

As sa.engine supports the ability to create and deploy neural net models that can be used in analytical models.

Should you already have existing analytical models implemented in standard C, Java, etc., then can easily be plugged in and re-used and augmented through our tools.

For further information on our proprietary OSQL domain calculus analysis and query language, please contact us

Revolutionizing the Model Development Process

Traditional/common model development (top circle) for data analysis is a centralized and slow process, usually taking hours/days/weeks to validate. Our technology accelerates and evolves this process, reducing validation to minutes (bottom circle):

  • Models can be developed, validated, and refined quickly
  • Analysts can experiment with data directly on equipment
  • Very fast deployment time

This provides better knowledge of equipment, better models to predict behavior and ultimately cost savings.

Deep Learning Everywhere

sa.eninge allows for building, training, and running neural networks everywhere.

Our very lightweight neural network builder, trainer and executor can be used on both mobile and edge devices as well as on clusters and workstations, using the same architecture everywhere.

Neural networks can be sent between all kinds of computers with no dependence on other software.

Our software is interoperabe with industry standard Google’s TensorFlow. This is done by converting centrally trained Tensorflow neural networks to our neural network module for pre- and post-processing, execution, and training anywhere.


Our software can be used in orchestration at several levels. For instance, our sa.engine may be deployed locally on multiple sensors on a single machine, reading and anlyzing all devices individually. To determine what is happening to machine as a whole, sa.engine may be orchestrating the devices via an installation on the microcontroller or similar device located on the machine. Finally, to orchestrate the data streams across multiple machines, sa.engine is installed centrally on a cloud solution, allowing it to manage streams across multiple machines.