How it works
Easy steps to build, train and tune your model in the browser-based tool

Drag and drop objects to build your model.
Connect the components and add layers.

Click on components to change hyperparameters or the code.

See all docs >

Run your model to train it, and watch it perform.

Export model for inference.

Start building for free today >
Building and training a model is a process
We’ve made it easier. And explaining your model - we’ve made that easier too.

Fast modeling

Drag and drop interface, with per-object previews

Use the GUI, custom code editor or Notebook

Every component/layer is programmable

Increased transparency

Instant feedback into your model’s performance

Real-time analytics in every operation and variable

Granular previews of output from each model component


Customizable environment and statistics dashboard

Local, on-prem or cloud-based version

Execute any custom Python code and export a fully-trained TensorFlow model


A Look at PerceptiLabs’ Browser-Based Architecture


MLOps: An "Ops" Just for Machine Learning


PerceptiLabs named to the CB Insights AI 100