aliu-airobot/eseilane

ESEILANE is a graph database purpose-built for AI applications that use language models. A knowledge graph is a way to store information as a network of connected facts: you store entities (like "Artificial Intelligence" or "GraphRAG") as nodes and the relationships between them as edges. ESEILANE is designed to query these graphs very fast, claiming sub-millisecond response times for most operations. The main use case is GraphRAG, short for Graph-based Retrieval-Augmented Generation. When you ask a language model a question, it can produce incorrect answers because it only knows what was in its training data. GraphRAG improves this by first retrieving relevant facts from a structured knowledge graph and passing those facts to the model as context before it responds. ESEILANE is built specifically for this workflow. Technically, the engine stores graph data using sparse matrix algebra via a library called GraphBLAS, which is why queries are faster than traditional graph databases. The core is written in Rust for performance and memory safety. You can query the graph using OpenCypher, a standard query language similar to SQL but designed for graph data. Setup is straightforward: start the database with Docker in one command, or install the Python package with pip. Python and TypeScript SDKs are both available. The GraphRAG integration lets you feed text into the graph, then ask questions that the engine answers by combining graph context with an LLM of your choice. Benchmarks in the README show ESEILANE 3x to 7x faster than Neo4j and ArangoDB on common operations. The license is Apache 2.0.