schema · evergreen
The vanilla RAG pipeline
Two separate tracks: the offline index build and the query path that runs on every request.
Vanilla RAG is two separate pipelines, and not confusing them is half the job: one runs offline and builds the index, the other runs on every request.
Rendering diagram…
Sourcedocs, tickets
Chunksplit + overlap
Embedvector
Storeindex
Query
Retrievetop-k
Promptcontext + question
Answerwith sources
Everything vanilla does not have
No router, no multi-step planning, no feedback. The model retrieves once and answers once. The real work is setting the contract between retrieval and generation: what you hand the model, and what you expect back.