Skip to content

Concepts ​

The mental model behind index4.ai's ingestion pipeline, hybrid retrieval, and prompt-ready output.

In this section ​

  • Ingestion What POST /ingest does and how a document becomes searchable.
  • Retrieval Hybrid search internals — min_relevance, max_tokens, document scoping.
  • Retrieval tiers The two per-query tiers — standard and advanced — and how they are billed.
  • RetrievalContext format The XML and Markdown grammar index4.ai returns, sandwich ordering, citations.
  • Document management List, inspect, and delete documents without touching the vector store.