Skip to main content

Guides

Welcome to the section of practical guides for using the Compressa platform.

Available Guides​

Basic RAG in 15 Minutes​

Step-by-step guide to creating a complete RAG pipeline using current information from external sources. You'll learn how to upload articles, create embeddings, set up semantic search, and connect an LLM model to generate relevant answers.

Detailed guide on working with text embeddings and creating semantic search that goes beyond standard keyword search. The guide includes loading a dataset with questions, converting text to numerical embeddings, setting up search, and visualizing results.

Improving Search Accuracy with Rerank​

Practical guide on using a two-stage approach with reranking to improve search result quality. You'll learn how the Compressa Rerank model prioritizes found passages for the user's query and improves result relevance.

Advanced Chunking for RAG Improvement​

Guide on optimizing the document splitting process into chunks using Compressa ETL, which is critical for RAG system efficiency.