Jina Supabase RAG FAQs

Question 1

How does Jina Supabase RAG discover and extract content?

Accepted Answer

It uses a smart URL discovery method that prioritizes sitemaps and falls back to recursive crawling via Crawl4AI. For content extraction, it employs a hybrid approach, primarily using Jina AI's fast reader API and Crawl4AI as a robust fallback for complex pages.

Question 2

What is Jina Supabase RAG?

Accepted Answer

Jina Supabase RAG is a focused server tool designed to crawl documentation websites and index their content, including text chunks and OpenAI vector embeddings, into a Supabase pgvector database. This enables efficient retrieval for Retrieval-Augmented Generation (RAG) applications.

Question 3

What are the main benefits of using Jina Supabase RAG for documentation?

Accepted Answer

Jina Supabase RAG provides a lean and efficient solution for creating a RAG-ready knowledge base from documentation. Its benefits include fast content extraction via Jina AI, flexible URL discovery, intelligent text chunking, and reliable Supabase pgvector storage, making it ideal for rapid deployment of documentation-based RAG.

Question 4

Can I index multiple documentation sites and keep them separate?

Accepted Answer

Yes, Jina Supabase RAG offers multi-project support. You can index various documentation sites, and each will be isolated under a distinct project name within your Supabase database, ensuring data organization and clarity.

Question 5

What kind of embeddings and database does it use?

Accepted Answer

The tool generates OpenAI vector embeddings (specifically 'text-embedding-3-small') for the extracted text chunks. These embeddings, along with the original content, are stored in Supabase's pgvector extension, which is optimized for similarity search.

Jina Supabase RAG

概要

主な機能

ユースケース