ApeRAG is a production-ready Retrieval-Augmented Generation (RAG) platform designed for building sophisticated AI applications, knowledge graphs, and intelligent agents. It targets engineers and researchers seeking to leverage hybrid retrieval, multimodal indexing, and scalable Kubernetes deployments for autonomous reasoning across diverse knowledge bases. The platform offers enterprise-grade management features and simplifies the creation of context-aware AI systems.

How It Works

ApeRAG integrates Graph RAG, vector search, and full-text search capabilities, augmented by AI agents and multimodal indexing. Its core approach utilizes a hybrid retrieval engine combining vector, full-text, graph, summary, and vision-based search methods. A key innovation is its deeply modified LightRAG implementation, featuring advanced entity normalization for cleaner knowledge graphs and improved relational understanding. The platform also supports Model Context Protocol (MCP) for seamless AI assistant integration and leverages MinerU for advanced document parsing, with optional GPU acceleration for complex content.

Quick Start & Requirements

The easiest way to start ApeRAG locally is via Docker Compose. Minimum system requirements include a CPU with at least 2 Cores and 4 GiB of RAM, along with Docker and Docker Compose installed. The setup involves cloning the repository, copying an environment template, and running docker-compose up -d --pull always. Users can access the Web Interface at http://localhost:3000/web/ and API Documentation at http://localhost:8000/docs. A live demo is also available.

Highlighted Details

• Advanced Indexing: Supports Vector, Full-text, Graph, Summary, and Vision indexes for multi-dimensional understanding.
• AI Agents & MCP: Features integrated AI agents with Model Context Protocol (MCP) support for autonomous knowledge base interaction and web search.
• Enhanced Graph RAG: Utilizes a modified LightRAG implementation with advanced entity normalization for improved knowledge graph accuracy.
• Multimodal & Parsing: Offers complete multimodal document processing, including vision capabilities, and integrates MinerU for advanced parsing of complex documents, tables, and formulas, with optional GPU acceleration.
• Hybrid Retrieval: Employs a sophisticated engine combining Graph RAG, vector, full-text, summary, and vision search methods.
• Production Deployment: Provides Kubernetes support via Helm charts and KubeBlocks for scalable, high-availability deployments of necessary databases (PostgreSQL, Redis, Qdrant, Elasticsearch, Neo4j).
• Enterprise Features: Includes audit logging, LLM management, graph visualization, and comprehensive document/agent workflow management.

Maintenance & Community

ApeRAG is actively developed, with community support available via Discord and Feishu. Links to community channels and contribution guides are provided within the repository.

Licensing & Compatibility

ApeRAG is licensed under the Apache License 2.0, which generally permits commercial use and integration with closed-source projects, subject to the license terms.

Limitations & Caveats

The provided README does not explicitly detail limitations, known bugs, or alpha status. While local deployment via Docker Compose is straightforward, production deployment requires a Kubernetes cluster and careful configuration of database services. The advanced document parsing with GPU acceleration is recommended, suggesting potential performance differences for complex documents without it.