Skip to main content

Super.AI: documentation and API reference

Super.AI is a data-labelling solution for companies looking to train or make use of machine learning models.

Read the documentationRead the API reference

I created the entirety of the super.AI documentation from scratch: nearly 100 pages, including getting started guides, how-to guides, conceptual introductions, and an API reference covering around 30 endpoints. I also made product feature videos to illustrate the value and use of new features upon release. The documentation went through several iterations before arriving at the configuration it had when I left, with ReadMe as the tooling.

Content layout​

I divided the content up conceptually, following the Diátaxis framework, with getting started guides designed to introduce newcomers to the product, how-to guides for users looking to achieve certain end goals, concepts for users to understand better how the product fits together and functions, and the API reference for the real nuts and bolts backend integration stuff.

The super.AI documentation landing page, with cards for getting started, how-to guides, concepts, and the API reference

Getting started​

I produced four getting started guides in total, one for first interacting with the product via the product store and dashboard, and an additional three for each method of interacting with the API (cURL, Python, and the CLI). All contain numbered step-by-step instructions alongside relevant code samples.

The Getting started with Python guide, showing a numbered step with a Python code sample

How-to guides​

These step-by-step instructions operate like recipes that direct users to specific end goals. These guides cover everything the user can achieve through the dashboard. Written content is bolstered throughout by screenshots and GIFs showing an overview of the process described in the instructions.

The How to label data yourself guide, showing a callout and numbered steps

Concepts​

The concept pages introduce the key parts of the product at a fairly abstract level to allow users to better understand the underlying machinery and make better use of the product as a result. By understanding the product concepts, users will be better equipped to make use of the more advanced how-to guides and the API reference.

The AI compiler concept page, with a diagram of how inputs are split into tasks and recombined

API reference​

The API reference provides nitty gritty coverage of all the publicly accessible API endpoints the product has available. This allows users to integrate super.AI with their own solutions or external databases. Every endpoint is documented alongside all its required and optional parameters and code examples (for calls and responses).

The API reference page for the Upload ground truth data endpoint, with parameters and example requests and responses