snippet · growing

The reference baseline stack

Express + LangChain + Chroma + Ollama. The zero point every measurement compares against.

The reference baseline — a working, plain, deliberately unclever vanilla RAG service. This is the zero point every measurement is compared against.

API & orchestrationTypeScript + Express

RAG integrationLangChain

Vector storeChroma, local

Embedding & generationOllama, temperature 0

ConfigurationZod + YAML

one process, three dependencies

Two tracks

track 1 — offline index build

Read JSON

FilterpdfText

TextSplitter

Metadata

Write to Chroma

track 2 — POST /rag/query

Zod validate

similaritySearchq, k

Format chunks

ChatOllama

JSON response

temperature = 0 is not an accident: during evals the variance in an answer should come from retrieval quality, not from sampling.