What is NLP enrichment?

After you commit data, Curiosity automatically processes text fields through NLP pipelines. Enrichment turns unstructured text into structured graph data.

---

Three things enrichment does:

1. Extract — find meaningful spans in free text

"The MBA-2024 battery drains overnight" → entity: MBA-2024, type: Product

2. Link — connect extracted mentions to existing graph nodes

MBA-2024 mention → edge _Mentions → Device node with key MBA-2024

3. Embed — generate dense vectors for semantic search

ticket body → 512-dimensional vector → stored in vector index

---

The result: your graph grows richer automatically as data flows in. Users can facet by extracted entities. The search engine finds semantically similar content. The AI layer has grounding material to work with.

You configure pipelines once. They apply to every node committed, past and future (after a reparse).

→ NLP overview