Get started | Umnai | Documentation

Use the UMNAI platform to turn source data into deployed Hybrid Intelligence models. Work through the web interface or use the REST API to integrate predictions and explanations into your applications.

How the platform works

Most workflows move through the same path:

Connect data

Grant the platform access to source data in your storage environment.

Ingest data into a datasource

Convert source files into a validated datasource managed by the platform.

Onboard the datasource into a dataset

Transform the datasource into a training-ready dataset.

Train a model

Train a model from the dataset.

Deploy the model

Create a deployed model that can serve predictions and explanations.

Query predictions and explanations

Explore the deployed model’s predictions and the explanation data behind them. Use views to inspect model behaviour, evaluate performance and understand individual results.

Get access

UMNAI sets up your organisation and account, then sends you an email invitation. Follow the invitation to create a password, then sign in to the web interface with your email address and password.

If you plan to use the REST API, sign in to the UI first to create your API credentials.

Choose your guide

Web app guide

Prepare data, train and deploy models, and explore explanations through the web interface.

REST API guide

Start with an end-to-end quickstart, then build and automate workflows with the REST API.

For the terminology behind both workflows, see Concepts.