surfaceextractiontriples
Extractionscalar · returns json

TRIPLES

Extract (subject, predicate, object) knowledge graph triples from text

Per-row — runs once for each row.

extractionllmtext

Syntax

TRIPLES({{ text }})
TRIPLES({{ text }}, '{{ focus }}')
{{ text }} TRIPLES

Arguments

nametypedescription
textVARCHAR
focus(optional)VARCHARFocus area: 'people', 'organizations', 'decisions', 'events', 'all'

About

Extract knowledge graph triples (subject, predicate, object) from text. Turns unstructured text into structured knowledge that can be queried, joined, and aggregated using standard SQL. Each triple represents a fact or relationship discovered in the text. Perfect for: - Building knowledge graphs from documents - Extracting relationships from emails/communications - Creating queryable facts from unstructured data - Entity relationship mapping - Information extraction at scale

Examples

Extracts employment and reporting relationships

SELECT
  triples (
    'John Smith works at Acme Corp as a software engineer. He reports to Jane Doe.'
  )

Extracts event information

SELECT
  triples (
    'The meeting on Tuesday was cancelled. Bob will reschedule for next week.'
  )

Nearby rabbit holes

same domain
Climb back to The Looking Glass