Repolode
Submit
apecloud

apecloud/ApeRAG

Production-ready RAG platform combining Graph RAG, vector and full-text search, multimodal parsing, and AI agents, deployable via Docker Compose or Kubernetes.

View on GitHub
Stars
1,322
Forks
147
Open issues
37
Last commit
last month

What is it?

ApeRAG is a retrieval-augmented generation platform for building a knowledge base over your own documents. It combines five index types (vector, full-text, graph, summary, and vision) so queries can be answered using several retrieval methods at once. Its graph layer is a modified version of LightRAG with entity merging to keep knowledge graphs cleaner.

The platform handles multimodal documents, including images, charts and tables, and offers an optional MinerU-based parsing service with GPU support for complex PDFs and formulas. Built-in AI agents can pick relevant collections, search content and run web searches. An MCP server lets AI assistants such as those supporting the Model Context Protocol query collections directly.

For local use, ApeRAG starts with Docker Compose. For production, a Helm chart deploys the app on Kubernetes, with KubeBlocks optionally provisioning PostgreSQL, Redis, Qdrant, Elasticsearch and Neo4j. The backend is FastAPI with a React frontend and Celery for asynchronous tasks.

Key features

  • 01Five index types: vector, full-text, graph, summary and vision
  • 02Entity normalization for graph RAG built on a modified LightRAG
  • 03MCP server for querying knowledge collections from AI assistants
  • 04Multimodal document parsing with optional MinerU and GPU acceleration
  • 05Helm chart and KubeBlocks scripts for Kubernetes deployment

Use cases

  • Building a private question-answering system over internal documents
  • Connecting an AI assistant to a team knowledge base through MCP
  • Parsing technical PDFs with tables and formulas into a searchable graph
  • Running a self-hosted RAG service on Kubernetes with managed databases

Quick start

  1. 1

    Clone the repository

    git clone https://github.com/apecloud/ApeRAG.git
  2. 2

    Enter the project directory

    cd ApeRAG
  3. 3

    Copy the environment template to .env

    cp envs/env.template .env
  4. 4

    Start the services with Docker Compose

    docker-compose up -d --pull always

Things to know

  • Requires Docker and Docker Compose, with at least 2 CPU cores and 4 GiB RAM.
  • Full Kubernetes deployment needs a cluster and, by default, the doc-ray service needs 4+ CPU cores and 8 GB+ RAM.

Topics

#rag#graph-rag#knowledge-graph#vector-search#mcp#ai-agents#kubernetes#llm#agents#context-engineering

FAQ

What is ApeRAG?
Production-ready RAG platform combining Graph RAG, vector and full-text search, multimodal parsing, and AI agents, deployable via Docker Compose or Kubernetes. ApeRAG is a retrieval-augmented generation platform for building a knowledge base over your own documents. It combines five index types (vector, full-text, graph, summary, and vision) so queries can be answered using several retrieval methods at once. Its graph layer is a modified version of LightRAG with entity merging to keep knowledge graphs cleaner.
Is ApeRAG free to use? What license does it have?
ApeRAG is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is ApeRAG still maintained?
Updated within 6 months. The last commit to ApeRAG was on Sep 12, 2026.
What can I use ApeRAG for?
Building a private question-answering system over internal documents Connecting an AI assistant to a team knowledge base through MCP Parsing technical PDFs with tables and formulas into a searchable graph Running a self-hosted RAG service on Kubernetes with managed databases
How do I install ApeRAG?
Clone the repository: git clone https://github.com/apecloud/ApeRAG.git Enter the project directory: cd ApeRAG Copy the environment template to .env: cp envs/env.template .env Start the services with Docker Compose: docker-compose up -d --pull always
Where can I get ApeRAG?
The source code and installation instructions are on GitHub: https://github.com/apecloud/ApeRAG

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects