Scailable.

Deploy your ML & AI models.
Instantly. Anywhere.

Convert AI & ML models to WebAssembly.
Deploy your models from Python or R.
Super fast. Super easy.

Sign up and get started!

Sign up and get started!

Scailable converts your AI and ML models to WebAssembly. The resulting optimized binaries run almost anywhere. Sign up, and deploy your models with just one line of code!

Check our terms and conditions here.
Already have an account?
Log in here.
Forgot your password? Reset it here.

Which models are we talking about?

Scailable allows you to transpile your fitted models to tiny, safe and efficient WebAssembly binaries through our sclblpy, our upcoming R package, or, for example, through uploading your (ONNX) model at admin.sclbl.net.

models

Python deployment

Scikit-learn
Stats
models
Xgboost

PyTorch (ONNX)
Tensorflow (ONNX)
KERAS (ONNX)
CoreML (ONNX)


R deployment

Bayesian additive regression trees

PyTorch (ONNX)
Tensorflow (ONNX)
KERAS (ONNX)
CoreML (ONNX)


Direct Upload

Compile your Rust, C or C++ based models to Wasm.

As long as they conform to our basic Sclbl Application Binary Interface (ABI), you can deploy them to any of our runtimes.
 

  Available
  Under active development



Where do you want to run your model?

When Scailable has converted your ML or AI model to a small, safe, fast and efficient WebAssembly executable, you will now be able to run it almost anywhere by deploying it to one of our minimal Sclbl runtimes. During our beta, we will be running your models on our own cloud servers. But our edge, browser, bare metal and other cloud runtimes are nearing completion!

models

In the cloud

Our Sclbl cloud
On premise

AWS
Azure
Google Cloud
Your server


In a browser

Chrome
Safari
Firefox
Edge
and many more..


Mobile, IoT, Edge

MCU's such as ESP32
Raspberry Pi's

Routers running OpenWRT
Google Android
Apple IOS
Your smart fridge
and many more..


  Available
  Under active development


 

Explore our demos.

Often, actions speak louder than words. So, we put up a number of demo's to showcase what Scailable can do for you.

With Scailable, your can deploy your deep learning models anywhere. Run your model on an IoT device. Or within your browser. For example like this CIFAR ResNet image recognition model. The choice is yours!
Scailable supports running complex models (such at Bayesian Additive Regression Trees) in the cloud, on the edge, or on a browser. Check how we can flexibly generate posterior predictives for automatic property valuation models, anytime, anywhere.
Sometimes what you need is performance. We provide nothing less. The inferences from our models are extremely fast; check out some of our benchmarks. Or play with our interactive MNIST number recognition model. It updates its estimates as fast as you can draw!
 

Take Scailable for a spin!

Want to get started as soon as possible? Sign up, try one of our tutorials, and see how easy it is to get started with Scailable!

A simple linear regression example. Here, we create a basic regression model using simulated data; give it a spin! This simple front-end supports our getting started tutorial: Scailable 101: Getting started.
Scailable also supports the deployment of complex deep learning models. Learn how to convert a ResNet PyTorch model to WebAssembly by exporting it to ONNX and running it through Scailable's ONNX toolchain.
Learn how to consume Scailable tasks on the web using our sclbl-webnode, which allows you to control whether to run tasks locally in the browser or through our RESTful API on our Scailable servers.
Would you like to deploy your own C or C++ based model using one of our runtimes? In this tutorial we explain how you create your own WebAssembly executable and upload it to Scailable to make it available as a REST endpoint.
 

About us.

It is estimated that up to 80 percent of AI and ML models never make it into production. Scailable was founded in 2019 by Robin van Emden and Maurits Kaptein to help bridge this gap.

The current team consists of a diverse group of people with backgrounds in coding, business and academia. Our shared goal is to move responsible AI to production to help it deliver on its many promises.

Maurits Kaptein
Maurits Kaptein
Robin van Emden
Arjan Haring
Arjan Haring
Arjan van den Born
Arjan van den Born
Davide Iannuzzi
Davide Iannuzzi
Fleur Hasaart
Fleur Hasaart
Petri Parvinen
Petri Parvinen
 
 
 
 
Powered by WebAssembly