Topic Clusters
Automatic grouping of related material across your corpus by thematic similarity — documents, conversation segments, extracted outputs, and notes that share concepts or subjects are clustered together without manual tagging.
Core Capabilities
| Emergent |
Dynamic |
Cross-Source |
| Clusters form from the content itself — no manual tagging or folder organisation required. |
Clusters evolve as new material enters the corpus, splitting and merging naturally. |
Documents, conversations, and extractions land in the same cluster when they share themes. |
The Problem We Solve
"My folders don't reflect how I actually think"
Folder hierarchies impose a single, rigid structure on material that belongs in multiple categories. Topic Clusters allow content to appear in every cluster where it's relevant — no duplication, no forced hierarchy.
"I organise things but can never maintain it"
Manual tagging is inconsistent and incomplete. You tag diligently for a week, then stop. Topic Clusters are automatic and continuous — they maintain themselves as your corpus grows.
"I know these ideas connect but I can't see how"
Two conversations from different months, a document from a different project, and a code snippet from a tutorial all relate to the same underlying concept. Topic Clusters surface these connections that manual organisation misses.
How It Works
- Analyse — Every source entering the corpus is analysed for thematic content using n-gram patterns, structural similarity, and semantic closeness
- Cluster — Sources with sufficient thematic overlap are grouped into clusters automatically
- Evolve — As new material enters, clusters grow, split, or merge based on the evolving landscape of themes
- Navigate — Browse the cluster hierarchy from broad themes to narrow sub-topics, or use clusters as filters in the Corpus Index
- Connect — Clusters with strong inter-connections surface on the Knowledge Graph as structural relationships
What We Deliver
- Automatic clustering based on semantic similarity, not just keyword matching
- Dynamic cluster evolution — clusters grow, split, and merge as new content arrives
- Cross-source clustering — documents, conversations, and extractions in the same cluster
- Cluster hierarchy — sub-clusters for granular thematic navigation
- Cluster-level analytics showing coverage, density, and temporal development
- Inter-cluster connections showing how different themes relate
- Cluster-based filtering across the Corpus Index and Source Browser
Integration with Other Features
- Corpus Index — Filter search results by cluster membership
- Node Graph — Clusters appear as navigable regions on the visual graph
- Source Browser — Each source shows its cluster membership in the contextual sidebar
- Entity Links — Entities spanning multiple clusters highlight thematic bridges
Related
→ Explore All Features