Skip to content

Build

Build (/build) is the hub for creating and populating knowledge graphs. It has three tabs:

Tab What it does
Knowledge Graphs Create, view, and manage your graphs
Ontology Define the schema (see Ontology Studio)
Ingest Load documents and data into a graph

Knowledge Graphs

The default tab lists your knowledge graphs with their status. A graph must be in the ready state before you can attach an ontology or query it. Create a new graph here, then move to Ingest to populate it.

Ontology

Define node types, relationship types, and properties for a graph. A schema makes extraction produce typed, validated entities. Full details in Ontology Studio.

Ingest

Load data into a graph. Graffold ingests multiple source types:

  • PDFs — including scientific papers (Nougat) and scanned docs (vision OCR)
  • CSV / Excel — tabular data
  • REST APIs — any API endpoint becomes a connector
  • Parquet — portable columnar data

Ingestion runs the pipeline: chunk → extract entities → resolve duplicates → build relationships → write to the graph. See Ingestion → graffold-ingest and Ingestion → Ingest API for the full pipeline and programmatic access.

Typical flow

1. Build → Knowledge Graphs → create a graph
2. Build → Ontology → define your schema (optional but recommended)
3. Build → Ingest → load your documents
4. Explore → query the result