What is Cluster?

Cluster is a research synthesis platform designed for UX researchers, product teams, and anyone who conducts qualitative research. It helps you:

1. Annotate research recordings, transcripts, and documents
2. Organize findings into thematic clusters
3. Synthesize insights backed by evidence
4. Export everything in open, portable formats

Core Concepts

Annotations

An annotation is a connection between your commentary and a specific piece of source material. In Cluster, annotations follow the W3C Web Annotation Data Model, which means:

• Highlights — Select text in a transcript or a time range in a video
• Tags — Apply taxonomy tags to categorize findings
• Comments — Add notes and observations
• Links — Connect related moments across documents

{
  "@context": "http://www.w3.org/ns/anno.jsonld",
  "type": "Annotation",
  "motivation": "highlighting",
  "target": {
    "source": "https://your-sharepoint.com/interview-p05.vtt",
    "selector": {
      "type": "TextQuoteSelector",
      "exact": "I had no idea what to do after I created my account"
    }
  }
}

Clusters

A cluster is a group of related annotations. Use clusters to:

• Group pain points by theme
• Organize feature requests
• Build affinity maps
• Prepare for synthesis sessions

Clusters appear on a visual canvas where you can drag, arrange, and connect annotations.

Insights

An insight is an evidence-backed claim. Unlike annotations (which are observations), insights are conclusions:

• "Users experience confusion immediately after signup"
• "The onboarding flow lacks clear next steps"
• "Power users want keyboard shortcuts"

Each insight links to the annotations that support it, creating a traceable chain from raw data to recommendations.

How Data Flows

┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐
│   Source Files  │────▶│   Annotations   │────▶│    Insights     │
│   (SharePoint)  │     │   (PostgreSQL)  │     │   (PostgreSQL)  │
└─────────────────┘     └─────────────────┘     └─────────────────┘
         │                      │                        │
         │                      │                        │
         ▼                      ▼                        ▼
   Files stay in         Annotations stored        Insights link to
   your cloud            as W3C JSON-LD           supporting evidence

Key insight: Your source files (videos, transcripts, documents) never leave your cloud storage. Cluster only stores references and annotations.

Supported File Types

Type	Extensions	Features
Video	.mp4, .webm, .mov	Time-based annotations, clip creation
Audio	.mp3, .wav, .m4a	Time-based annotations
Transcripts	.vtt, .srt, .txt	Text highlighting, speaker labels
Documents	.docx, .pdf	Text highlighting

Who Uses Cluster?

UX Researchers

Primary users who conduct interviews, usability tests, and synthesis sessions. Cluster replaces or augments tools like Dovetail, Marvin, or Condens.

Product Managers

Secondary researchers who review research findings, create insights, and connect research to product decisions.

Research Operations

Teams responsible for research infrastructure, data governance, and tool administration. Cluster's self-hosting model gives ResearchOps full control.

Agencies & Consultants

Research practitioners who need to hand off data to clients. W3C-standard export means clients aren't locked into Cluster.

Next Steps

• What Cluster is Not — Understand the boundaries
• Hosting Options — Choose your deployment model
• Architecture Overview — Technical deep-dive