Coming Soon
Deploy with Docker
AI Runtime and Management for Control Systems
Maximize profits, boost energy efficiency, and increase process uptime for heavy industry with KoiosLT, our AI runtime system.
How KoiosLT Works
Easily integrate AI Models into your existing Control System for real-time optimization results in just a few clicks.
Execute
KoiosLT software installs with most off the shelf hardware or virtual appliance setups and can deploy AI models on any controls network. KoiosLT can securely exists either connected to the internet or completely air gapped.
Standard industrial protocol integrations
Real-time AI engine
High availability and redundancy
Trending capability
Offline patching capabilities
SNMP integration
Services
Flexible AI Options
Our flexible AI options can adapt to what you need.
Custom AI Modeling
Our data scientists will work with you to create exactly what you need.
Front-end feasibility studies
Build and validate models
Manage models and data
Already have an AI Model?
Have your own AI Model? With KoiosLT there is no vendor lock-in. You can deploy your models with KoiosLT, and the time & effort invested engineering your models is yours to keep.
Flexible AI engine
Common protocols
Multiple add-ons
Industries Served
The following industries can be assisted by our AI models.
Oil & Gas
Wastewater
Utilities
Chemical
Agriculture
Renewables
Commercial
Manufacturing
Pharma
The Ai-controller helped the hydrotreater process unit operate with more stability producing better heat distribution through the exchangers, less flow valve cycling and longer run times. We are already looking at other processes to evaluate the software.
Bobby Newsome
Manager, Transmix
Reliable Runtime & Peace of Mind
Describe some metrics that your customers care about.
30,000
hours of runtime since 1st deployment
600%
typical return on investments
< 1%
hours added to personnel management workloads
KoiosLT System Requirements
Simple, flexible, seamless
Real-time AI advanced control systems that work seamlessly with most off the shelf hardware. KoiosLT makes developing and deploying trained models on any controls network a real-time possibility.
CPU
Supports x64 and x32bit Architectures
Operating System
Linux
Memory
8 GB RAM
Storage
250 GB, preferred SSD or NVMe
Communication
2 x 1GB Ethernet Ports
Note: The number of models that can concurrently execute on the engine is dependent on the model complexity and the hardware selection see the below Matrix for estimated engine performance
Get started
Need answers?
Frequently asked questions
  • KoiosLT is a software platform developed by Ai-Ops, Inc, that comes equipped with a real-time engine that executes tflite files. The KoiosLT comes equipped with secure, modern industrial APIs for system integration options and can easily connect to many existing controls networks. The KoiosLT comes with all the tools required for a control system engineer to deploy, manage, and monitor AI models on controls networks.
  • The KoiosLT uses tensorflow lite as its predict engine's backend framework. The KoioisLT will run any models file that is uploaded in .tflite format, and was trained using two-dimensional, timeseries data. The predict engine comes equipped with its own critical and secure database for holding the required istory to execute the models with the highest resolution and perfomance. The predict engine can handle many flexible execution rates, from the milliesecond range to days and months.
  • The KoiosLT is designed to run 100's of models concurrently. The amount of models and IO is dependant on the hardware or VM resource selection. Anything from a RaspberryPi, high powered physical server, or a virtual machine can be used to deploy the KoiosLT software.