OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
RAG & Knowledge Systems

Vector Graph Rag

Vector Graph RAG is a Python library for building graph-augmented retrieval-augmented generation systems that encode entities and relations as vectors in Milvus and retrieve evidence for multi-hop questions.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Vector Graph Rag — OpenViST project profile in RAG & Knowledge Systems.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when an application needs multi-hop evidence retrieval from a knowledge base and you want a Milvus vector-search approach without a separate graph database.

Who it is for

  • Developers building knowledge-base question-answering applications.
  • Teams working with fact-dense material such as legal, finance, medical or literary content.
  • Practitioners comparing retrieval strategies for multi-hop questions.

Documented uses

  • Build a knowledge base from text passages and query it with factual questions.
  • Retrieve evidence that connects facts across multiple passages.
  • Use pre-extracted subject-relation-object triplets instead of relying on LLM triplet extraction.
  • Compare single-pass generative reranking with two-stage Jev relation and passage scoring.
Before you choose

Requirements & limits

The repository specifies Python 3.10 or later and an MIT license.

  • The retrieval design uses Milvus vector search and subgraph expansion rather than a separate graph database or iterative search loop.
  • A basic text-ingestion quick start reads OPENAI_API_KEY from the environment.
  • Optional installation extras are documented for PDF, DOCX and web-page loaders; HuggingFace, Ollama and Jina embedding integrations; and OpenTelemetry tracing.
  • The README reports Jev evaluation results on MuSiQue and 2Wiki; these are project-reported evaluations.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Build and query a text-based knowledge base

Create a Vector Graph RAG knowledge base from text passages and query it.

Before you start

  • Python 3.10 or later
  • Install the vector-graph-rag package with pip or uv
  • Set the OPENAI_API_KEY environment variable

Steps

  1. Install vector-graph-rag with pip or uv.

  2. Set OPENAI_API_KEY in the environment.

  3. Import VectorGraphRAG and create an instance.

  4. Pass text passages to rebuild_texts().

  5. Call query() with a question and read the result's answer.

Expected outcome

A query result containing an answer based on the rebuilt text knowledge base.

Guide limitations

  • This documented quick-start path uses an OpenAI API key.
  • The README describes Milvus Lite as the default; it does not require a separate graph database.
Documentation 1 ↗
Source-based practical guide

Build a knowledge base from pre-extracted triplets

Load passages with existing entity-relation triplets without using LLM triplet extraction.

Before you start

  • Python 3.10 or later
  • Install the vector-graph-rag package with pip or uv
  • Passages and their knowledge-graph triplets

Steps

  1. Create a VectorGraphRAG instance.

  2. Prepare passage records with triplets represented as subject, relation and object entries.

  3. Pass the records to rebuild_documents_with_triplets().

Expected outcome

The library rebuilds its knowledge base using the supplied passages and triplets.

Guide limitations

  • This route skips LLM extraction only when the triplets have already been prepared.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Retrieval-augmented applications

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
Vector Graph RagBuild a knowledge base from text passages and query it with factual questions.Linux, macOS, WindowsThe repository specifies Python 3.10 or later and an MIT license.
LangsearchCrawl website or local data for a knowledge base.CLI, Self-hosted, Server
Hybrid Search RagCombine document retrieval and language-model generation in a configurable workflow.
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
12 October 2026
Original website title
zilliztech/vector-graph-rag · Official homepage title · retained in its source language
How we review sourcesSuggest a correction