Clusters

What is a Cluster?

A cluster is a thematic grouping of signals. While signals capture individual moments, clusters organize those moments into meaningful arcs -- connecting signals by project, time period, location, theme, or any other organizing principle.

Clusters sit in the middle of the preservation hierarchy: Shape > Cluster > Signal.

Use Cases

Clusters are flexible and can represent many kinds of groupings:

• A project arc -- All signals related to a specific project or endeavor
• A time period -- Signals from a particular week, month, season, or era
• A location-based collection -- Signals captured in the same geographic area
• A theme -- Signals connected by subject matter, mood, or narrative thread

Cluster Analysis

When a cluster contains signals, AI can analyze the cluster as a whole. This cluster-level analysis looks across all member signals to identify:

• Common threads and recurring patterns
• The narrative arc formed by the signals together
• Relationships between signals that may not be obvious individually

Cluster analysis produces a unified view that is more than the sum of its parts.

Cluster Embeddings

Each cluster receives its own vector embedding based on its collective content. This enables semantic search at the arc level -- you can find clusters by meaning, not just by keyword, surfacing thematic connections across your entire archive.

The Clusters List View

The clusters list provides an overview of all clusters. You can browse, search, and filter clusters by name, date range, or signal count. Each cluster entry shows a summary of its contents and analysis status.