Data Collections: Overview

A Collection is a way to create a "fence" around subsets of your data, so you can work with just that data. Instead of asking Storytell to answer questions based on All Knowledge every time, you can define a smaller, more relevant subset of your data and save it as a named Collection. This gives you control over what the AI considers as inputs to its answers, helps reduce noise, and allows you to reuse the same scope across different chats and workflows.

Written By Patrick Intervalo

Last updated About 2 months ago

ℹ️ Think of a Data Collection as a “fence” you create around a subset of your data. Your entire Knowledge base is one large Collection by default, called All Knowledge. You can create smaller “fences” to chat against and use in workflows. Collections also allow you to see just the Concepts present in that fenced ring of data.

Why use Collections?

  • Stay focused: Avoid noise by working only with the Assets and Concepts relevant to your task.

  • Work faster: Save scopes you return to often instead of recreating them each time.

  • Collaborate clearly: Share Collections with teammates so everyone works from the same context.

  • Adapt easily: Create different Collections for research, strategy, planning, or reporting, and switch between them as needed.

ℹ️ Did you know that you can turn concepts into visuals like images, infographics, and word clouds with 1-click? Learn how to do it here.

How Collections Work in Storytell

  1. Create a Collection by using the filter icon in the default All Knowledge Collection. Select one or more labels you want to use to define your Collection.

  2. You can click the include all of filter text to modify it to include, any, exclude all, or exclude any.

  3. You can keep clicking the filter icon to select multiple filter sets, and you can choose to match all or match any of those filters.

  4. When you have crafted the data set you want for the Collection based on the label filters, choose Save Collection to save the Collection you’ve defined as a persistent data set to work with.

Examples of Collections:

Below is a simple Collection, Just Planets with just one label applied. Using @mentions you can scope this Collection to your chats, although in this case, you could just as easily at-mention the label in your chat and the effect would be the same.

click images to enlarge

This Collection, Planet or Star Data is more sophisticated. It is defined to match all content that contains either the Planet Data label or the Star Data label.

In this case, Using @mentions, if you were to at-mention the Planet or Star Data Collection in your chat, you would be chatting with all the matching files below.

This Collection, Planet and Star Data is set to only match content that contains both labels. In this case, only one asset has both Planet Data and Star Data applied. (If any future assets were labeled with both labels, they would automatically show up in this Collection).

This Collection, Planet or Star data (but not images) contains multiple filters. You can set multiple filters by clicking the filter icon again after you create your first filter. This Collection is set to exclude any assets labeled as images — even if they have one of the other labels applied.

Other things to keep in mind:

  • Scope control: Mention a Collections by name to limit prompts to be run on just the data in that Collecction.

  • Knowledge Pane: Switch between Assets and Concepts within a Collection for different perspectives on your data

  • Projects: Collections defined in one project won’t exist in other projects. Collections, Labels and Assets are all defined per-project.

  • Coming soon: Automations that rerun prompts when the data in Collections changes, keeping insights continuously up to date.

Collections and the bigger picture

Collections connect with other Storytell features:

  • They fence your data based on usage of one or more Labels, which help you organize your Assets

  • You can see either a list of the assets in a Collection, or a graph of all the the Concepts in that Collection, letting you see extracted insights only from the scope that matters.

  • They integrate into Prompts and Actions, giving you precise control over the data Storytell draws from.

By combining these elements, Collections become more than organizational tools — they shape the way you interact with your knowledge.

Deleting a Collection

ℹ️ Think of a Collection as a “fence” you create around a subset of your data. When you delete a Collection, you’re just removing the fence, but you’re not doing anything to the data itself, or the labels you’ve placed on that data.

Removing a Collection will just remove that “fence” around your data. It will not affect or change the data itself, or the labels you’ve applied to your data.

To delete a Collection, click the trash can icon: