Skip to content

Hi Coactive team!

This is a sample of what your docs might look like on Starport, based on your public docs.

Take a look! Ask AI, search, the MCP server, and Markdown copies all work.

Starport is a free and open-source docs framework based on Starlight and maintained by Promptless. Promptless is the AI agent that automatically updates your customer-facing docs.

Every annual Promptless plan comes with white-glove migration to Starport, where we migrate the content, tune the result with you, and you own the repository so you're never locked in.

Book a 15-minute walkthrough

Sample migration of Coactive docs to Starport, prepared by PromptlessBook 15-minute call

Intelligent Search

Coactive’s Intelligent Search is designed to expedite and enhance content discoverability. This feature allows you to search through your content based on a textual description (e.g., “school of fish,” “A couple skiing as the sun is setting,” etc.) without the need for pre-populated tags or metadata. Users can search for people, locations, objects, events, actions, a combination of the above… and more.

The search experience lets users integrate a search query with custom metadata generated on the platform called “Coactive Concepts.” These tags can represent abstract ideas, objects, or actions specific to your enterprise’s unique taxonomy.

  1. Log into your Coactive account
  2. In the “Datasets” tab, select the dataset with your assets.
  3. At the bottom of the search bar, you’ll see that “Visual search” is the default selected option. Overview of the Coactive semantic search feature
  4. Enter your search query and click Enter or the arrow. Overview of the Coactive semantic search feature
  5. If you’d like to integrate one of your concepts, it’s as simple as using the concept drop-down and selecting your concept and hitting enter or the search arrow. Overview of the Coactive semantic search feature
  6. Want to narrow your results down even further? Select your concept and then use Intelligent Search to type in your query! Overview of the Coactive semantic search feature

“Query engineering” involves designing and optimizing visual search queries to retrieve the most relevant videos. However, results may sometimes not meet expectations due to how machine learning models interpret visual data.

  • For example, searching for “home interior” might return results that include home offices, kitchens, living rooms, etc., but require additional effort to find exactly what you’re looking for.
  • Adding detail, such as “a man working on a couch in his home”, provides more context and improves search accuracy.

Including more detail can help refine your queries and help you get your desired results quicker.

  1. Start Broad, Then Narrow Down: Begin with a broader query and gradually add more details.
    • Examples:
      • Basic: friendship
      • Medium: friends enjoying a meal together
      • Advanced: friends enjoying a meal together at night
  2. Review and Refine: After viewing the initial results, refine the query based on what was or wasn’t found. Continue to add more specific details.
  3. Use synonyms: Think of different words or phrases that mean the same thing or are related to your original query.
    • Examples: For a query like “mature adult,” also try out “elder” and “senior citizen”.

When searching your content, you might find yourself looking for something that isn’t visually distinct. Audio search allows you to leverage natural language queries to explore your content’s audio track, just as you would search a transcript (e.g. “I have a dream”, “Princess Diana killed in Paris car crash”, etc.), all without the need for pre-populated metadata.

Within our media use cases, we typically see audio search most helpful when users are looking to find a specific event that is less defined by visual elements, for example:

  • A specific quote
  • A contextual event (i.e. a description of an event or a headline)
  • … and more!

With Coactive’s audio search, you use quotes or natural language to pinpoint content based on what was spoken or reported audibly - no metadata needed.

  • In the dropdown under the search bar, select “Audio Search”, type your phrase in, and select enter or the arrow.

    Overview of the Coactive semantic search feature