Corpus & data pipeline
Ingest
Ingesting is the first-time intake of a source into the system - fetching it, parsing it, and adding it to the corpus.
Ingest is the pipeline step that brings a new source into the system for the first time: fetch the raw data, extract the useful parts, and store a structured record.
It is distinct from later refreshes. Ingest creates the record and its initial relationships; subsequent runs only update it.
In GraphCanon
When a GitHub repo is submitted, ingest fetches its metadata, generates a neutral summary and tags with an LLM, and proposes typed edges to existing tools - all in one pass.
See also
Related terms
Last reviewed 2026-07-09