RAG Pipeline

Retrieval-Augmented Generation · Notion · Chroma · Qdrant

The $0 AI Architecture Stack · 2026

What is the RAG Pipeline?

The RAG Pipeline gives the LLM external knowledge it wasn't trained on. Source documents (e.g. from Notion) are chunked, embedded, and stored in a vector database. At query time the most relevant chunks are retrieved and injected as context. Use Chroma for simple local storage or Qdrant for a fast, production-grade vector DB running locally.

Tools in this Layer

📔 Notion

Use Notion as a knowledge base — sync pages and databases as RAG source documents.

🌈 Chroma

Lightweight, embeddable vector store — perfect for prototypes and small datasets.

🟥 Qdrant

High-performance vector database with filtering, runs locally with zero cost.

Cost

$0 / free tier

Chroma and Qdrant are open-source; Notion has a generous free plan.

Explore the Full $0 AI Stack

1 Frontend Layer 2 Agent Orchestrator 3 RAG Pipeline 4 LLM Layer 5 Tool Use via MCP 6 Code Agent 7 Data Layer 8 Deployment Layer