This section describes the key concepts of the platform.


Each account using the platform is identified as an organisation. All resources you can manage and query belong to your organisation. Each organisation is identified by it’s name and the name is globally unique on the platform.


Datasets are collections of content meant to be explored together. Each organisation can have many datasets. Like organisations, a dataset is identified by its name. Each dataset must have a unique name within your organisation.

The dataset is an important concept because the platform is designed to work with content within the context of a collection of other content. A single item of content can belong to many datasets and so you can create multiple datasets to allow your content to be discovered in different contexts.

For example, if you have a large catalogue of content, you can create a single dataset containing everything to allow for it to be discovered as a whole. However you may also have several categories within your content and wish to also use the Visii platform to explore content within a specific category. In this case you can also create a dataset per category and decide which dataset to use depending on context.


Each image is referred to as a piece of content. As described above, a piece of content can belong to many datasets. When content is added to a dataset, we require a URL at which the content is hosted and a unique ID within that dataset. The platform will retrieve the content and process it. The content of the image is only used for the initial processing and once that is complete it is not stored.

After initial processing, whenever the Vision API is queried, results will refer to content by the unique ID specified by you. This is the ID for that content within the dataset currently being queried. You should then be able to retrieve the relevant details for the content from your own systems using that ID.