Skip to main content

Search Clusters

Use search in Clusters to find the groups of highlights that matter most, whether you are looking for a topic, a label, or a cluster someone on your team created.

Clusters search combines:

  • Keyword search for the words in a cluster title and description
  • AI-powered semantic search to find related clusters, even when they do not use the exact same wording
  • Filter-based search to narrow results by label, creator, and recording date

What search looks at​

Search on clusters is based on the cluster itself. In practice, that means search can use:

  • The cluster title
  • The cluster description
  • Cluster metadata such as labels, creator, and recording date

Clusters search is best for finding the right grouped theme or synthesis area first.

Start by typing a normal phrase, topic, or cluster name.

Examples:

  • pricing
  • onboarding friction
  • retention risks
  • enterprise requests

If you want to search for an exact phrase, put it in quotes.

Examples:

  • "retention risks"
  • "enterprise requests"

This is useful when you want a literal match instead of a semantic one. Semantic search matches based on meaning, so the exact words do not have to appear. Exact matching is useful when you are looking for product names, phrases you remember, or other specific wording.

For plain-text searches, results are shown using the cluster search sorting in NEXT. Depending on the selected view, clusters can be ranked by recency or by the number of linked highlights.

Use filters to narrow results​

You can combine a free-text search with filters, or use filters on their own.

FilterWhat it does
label:Returns clusters that include this label
creator:Returns clusters created by this person
date:Returns clusters created on the selected date or date range

Examples:

  • pricing label:Urgent
  • creator:Alice
  • date:Last 30 days

Filter tips​

  • Use label: when you want to browse clusters that share the same label.
  • Use date: to focus on clusters built from recent conversations.

Suggestions and autocomplete​

As you type, NEXT suggests filters and values to help you build your search faster.

For clusters, suggestions focus on:

  • Date ranges

When you start typing a specific filter such as label: or creator:, NEXT will suggest matching values for that filter.

Date suggestions support both quick ranges and natural language. For example, you may see options like:

  • Last 7 days
  • Last 30 days
  • Last 90 days
  • A parsed date range from something like date:December 2024

Once you already use a filter in the current search, NEXT avoids suggesting the exact same filter value again.

There can also be a short delay before new suggestions appear. Suggest search is cached for up to about 5 minutes, so recently added clusters, labels, or other updated suggestion values may take a few minutes to show up.

How to get better results​

  • Start broad, then add filters once you see the result set.
  • Search by topic first if you are not sure how a cluster was named.
  • Use quotes when you want to search for an exact phrase.
  • Use label: when your team uses labels to organize cluster types or workflows.
  • Use creator: when you want to find clusters created by a specific teammate.