A high-performance GraphRAG framework in Rust that transforms documents into knowledge graphs for superior retrieval and generation.
EdgeQuake is a high-performance GraphRAG framework written in Rust that transforms documents into intelligent knowledge graphs. It solves the problem of traditional RAG systems losing structural relationships by implementing the LightRAG algorithm, which extracts entities and relationships to enable multi-hop reasoning and thematic queries.
Developers and organizations building advanced document retrieval systems, AI agents, or knowledge management applications that require deep reasoning beyond simple semantic search.
Developers choose EdgeQuake for its combination of Rust's performance with graph-enhanced retrieval, offering superior query capabilities, production-ready features, and the ability to handle complex document types like scanned PDFs through vision LLMs.
EdegQuake 🌋 High-performance GraphRAG inspired from LightRag written in Rust; Transform documents into intelligent knowledge graphs for superior retrieval and generation
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Leverages Rust's async Tokio runtime and zero-copy operations to achieve sub-200ms query latency and handle thousands of concurrent users, as benchmarked in the README.
Implements the LightRAG algorithm with six query modes, enabling multi-hop reasoning and thematic analysis beyond traditional vector search.
Supports both text-based and vision LLM extraction for scanned documents, tables, and multi-column layouts, with embedded pdfium for zero-config setup as of v0.4.0.
Offers an OpenAPI 3.0 REST API with SSE streaming, health checks, and multi-tenant workspace isolation, making it suitable for enterprise deployment.
Requires PostgreSQL with pgvector and Apache AGE extensions, plus Docker for full-stack deployment, adding operational overhead compared to simpler RAG solutions.
Demands Rust toolchain for development and relies on external LLM providers for entity extraction and vision processing, which can increase costs and complexity.
As a version 0.9.1 project, it may undergo breaking changes and lacks the long-term stability assurance of a mature, production-grade release.