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

Minirag

Minirag is an open-source Python retrieval-augmented generation framework that combines document retrieval with small language models. Its documented approach uses heterogeneous graph indexing and lightweight topology-enhanced retrieval.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Minirag — 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 Minirag when the project calls for an open-source Python RAG framework focused on small language models, graph-based retrieval, and lightweight deployment patterns.

Who it is for

  • Developers building lightweight RAG prototypes
  • Teams exploring private-document question answering
  • Researchers testing small-model or on-device knowledge retrieval
  • Python users creating self-hosted documentation or research search tools

Documented uses

  • Index a document or benchmark dataset
  • Retrieve relevant knowledge for a question
  • Generate answers grounded in retrieved information
  • Prototype a small-model knowledge assistant
  • Reproduce the project’s documented RAG evaluations
Before you choose

Requirements & limits

The repository is publicly accessible, marked active in the supplied record, and released under the MIT license. The README documents installation from source and a package-based installation command, as well as API and Docker deployment documentation.

  • The system’s retrieval design is based on heterogeneous graph indexing rather than only a conventional vector-search description.
  • The README reports support for more than 10 heterogeneous graph databases, including Neo4j, PostgreSQL, and TiDB.
  • The project includes reproducible indexing and QA scripts plus the LiHua-World dataset.
  • The documented benchmark tables cover several small language models and GPT-4o-mini, but they are project-reported results.
  • The repository is Python-based and includes API and Docker deployment documentation.

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

Index a dataset and run the documented QA example

Prepare a dataset, create its index, and run the repository’s example question-answering workflow.

Before you start

  • A local copy of the MiniRAG repository.
  • A dataset placed in the repository’s dataset directory.
  • Python dependencies installed as documented by the project.

Steps

  1. Install from source with the command below, or use the package command shown in the README.

    Documented command
    pip install -e .

  2. Place the selected dataset in ./dataset/xxx; the repository includes LiHua-World under ./dataset/LiHua-World/data/.

  3. Run the command below to index the dataset.

    Documented command
    python ./reproduce/Step_0_index.py

  4. Run the command below to execute the documented QA stage.

    Documented command
    python ./reproduce/Step_1_QA.py

Expected outcome

The documented reproduction workflow indexes the selected dataset and proceeds to its question-answering stage.

Guide limitations

  • The README does not specify the model, infrastructure, or configuration needed for every dataset.
  • The supplied steps document the example workflow, not application-specific evaluation or production readiness.
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
MiniragIndex a document or benchmark datasetCloud, Self-hosted, ServerThe repository is publicly accessible, marked active in the supplied record, and released under the MIT license. The README documents installation from source and a package-based installation command, as well as API and Docker deployment documentation.
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
1 October 2026
Original website title
HKUDS/MiniRAG · Official homepage title · retained in its source language
How we review sourcesSuggest a correction