Index Refresh
Also called Index Refresh
Rebuilding or updating the search index so it reflects the latest source data.
Think of it like
Re-shelving the library after new books arrive — until you do, the card catalog points to gaps.
Example
A nightly job re-embeds changed documents and updates the vector index, so this morning’s edits become searchable by tomorrow.
How it actually works
When source documents change, the index built from them doesn’t update itself. Index refresh re-processes new or changed content — re-chunking, re-embedding, and upserting into the store. It can be a full rebuild (simple, expensive) or incremental (fast, fiddlier to get right). Cadence is a direct trade between compute cost and freshness; skip it and staleness sets in.
For product teams
The recurring maintenance that keeps answers current; skipping it is how a good system rots.
For engineers
Re-chunk and re-embed changed sources and upsert into the index; choose full vs. incremental by cost and change volume.
Related
- Freshness — What it maintains.
- Staleness — What it prevents.
- Embeddings — The step that regenerates vectors.
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome