Comment on page
Welcome to Scailable
Scailable provides a platform for creating and managing Edge AI solutions at scale. With Scailable, you can simply configure any supported edge device (such as a router, gateway, or IPC) to turn it into a "smart" device.
Smart devices can run advanced Artificial Intelligence (AI) and Machine Learning (ML) models on input data, such as a video stream, to turn the input into something meaningful, such as a count of the number of cars in the video.
These docs describe how you can setup and configure your own edge AI solutions on supported edge hardware.
Our docs are intended for our resellers and platform users who aim to create and manage edge AI solutions. If you are looking to purchase a solution directly, or if you are looking for use-cases, please visit us at www.scailable.net.
When reading the docs its good to keep in mind our general architecture which we display below:
Using Scailable, you can create Edge AI solutions, i.e., a solution that uses an AI or ML model running on an edge device.
The starting point for creating solutions is the Scailable AI manager, this is a bit of software running on an edge device (such as a gateway or IPC) that allows your to simply configure the AI solution you want to create by simply selecting an AI model and by configuring the device settings. Learn more about setting up the AI manger here.
The Scailable AI manager will take input from various sensors. At this point we support video imput (i.e., images coming from camera's.). Read more about the camera's we support here.
You can use the Scailable Platform (which you can read more about here) to remotely configure your solution and manage your solution at scale: i.e., if you have hundreds of devices, you can manage them in one go.
The core "magic" that goes into creating edge AI solutions is the AI model; the model effectively transforms the input (images) to the desired output (a count of the number of people in front of the camera, a "OK" / "NG" output for product inspection, the license plate of a car in front of the camera, or whether or not a person in front of the camera is wearing a helmet). We have a large list of off-the-shelf models available in our model library which allows users to simply configure new solutions. However, if you are a data scientist, you can also create your own models and upload them to our platform for your custom needs.
Finally, the data that is generated by the AI model is send to some application platfrom. We can log the data back to the Scailable platform, but we can also forward it to your destination of choice.
Note that all the actual AI/ML processing happens on the edge device. Thus, after configuring your device using the Scailable cloud, it can effectively be de-coupled from the world-wide-web to ensure fully remote operation.
Here we provide a short list of terms that pop up repeatedly in these docs and are good to know:
- The Scailable AI manager: The software running on the edge device that allows you to configure your edge AI solution. We simpy us "AI manager".
- A Scailable AI/ML model: We use this term somewhat loosely to any model definition, in Scailable portable model format (SPMF) that describes the transformation of input (images) to output (license plate, "OK"/"NG", etc.). This model can originally be a Deep Neural Network (AI), a simple classification model (ML), or just a traditional Vision pipeline. In the end, its just the logic running on the device to do the 'magic". We often simply use "AI model".
- The Scailable edge-ai management platform (or "Scailable platform"): Our cloud environment (which can be found at https://admin.sclbl.net) from which you can a) manage your devices and the AI models deployed to them, and b) add your own models to your personal AI model collections. We simply use "Platform".
- Edge device: We use this term somewhat loosely to refer to any device (camera, gateway, IPC, etc.) that runs the Scaialble AI manager. We simply use "Edge device".
- The Scailable data logger: The part of our platform devoted to logging (meta-) data originating from any edge device that has the Scailable AI manager installed.
Last modified 20d ago