Neo4j GraphRAG FAQs

Question 1

Does Neo4j GraphRAG support different embedding providers?

Accepted Answer

Yes, built on LiteLLM, Neo4j GraphRAG offers robust multi-provider embedding support. This allows you to utilize diverse AI models from providers like OpenAI, Azure, Bedrock, Cohere, and even local options such as Ollama, offering great flexibility for your embedding needs.

Question 2

How does Neo4j GraphRAG benefit LLM and RAG applications?

Accepted Answer

It empowers LLMs to intelligently query and retrieve information from your Neo4j graph database. By combining various search types with precise graph traversal, it provides highly relevant and contextually rich data, significantly improving the accuracy and depth of Retrieval-Augmented Generation (RAG) outputs.

Question 3

What is Neo4j GraphRAG?

Accepted Answer

Neo4j GraphRAG is an MCP server that significantly extends Neo4j's capabilities with advanced search features like vector, fulltext, and search-augmented Cypher queries, specifically designed to enhance Large Language Model (LLM) applications with graph data for Retrieval-Augmented Generation (RAG).

Question 4

What types of search can I perform with Neo4j GraphRAG?

Accepted Answer

You can perform semantic vector searches using Neo4j vector indexes, keyword fulltext searches with Lucene syntax on fulltext indexes, and execute complex Cypher queries that are dynamically augmented by either vector or fulltext search results. It also supports standard read-only Cypher queries for general graph traversal.

Neo4j GraphRAG

Neo4j GraphRAG

主な機能

ユースケース

主な機能

ユースケース