October 2025 Summaries

FalkorDB has been named a finalist in the Cyber Startup Challenge 2025, organized by the Cyber-Defence Campus of armasuisse Science and Technology, which focuses on innovative graph analytics technologies that can boost Switzerland's cyber defense capabilities. This multi-tenant graph database is noted for its impressive performance, delivering 10.5 times faster median latency than Neo4j for cloud security workloads, and is designed to map attack surfaces and threat relationships across various cloud security deployments like CNAPP, CSPM, and CIEM. By organizing security data as interconnected nodes and edges, FalkorDB enables real-time identification of potential attack paths, such as vulnerabilities connected through over-permissioned IAM roles. Supporting over 10,000 isolated tenant graphs, FalkorDB is particularly beneficial for large-scale applications, including SaaS and RAG systems, and offers open-source resources for developers interested in graph-based security analytics.

FalkorDB has released the @falkordb/langchain-ts package, which allows JavaScript and TypeScript applications to convert natural language questions into Cypher queries for graph databases, facilitating more intuitive data interactions without requiring users to write complex queries. Available on npm, this package integrates seamlessly with the LangChain ecosystem, enhancing AI capabilities by combining graph database relationships with language model features for more accurate responses. The integration leverages FalkorDB's low latency and is designed to fit into existing LangChain workflows, offering a straightforward setup through Docker for local development and providing type definitions via a TypeScript client. This release aims to simplify the user experience by automating the translation of natural language into graph database queries and is supported by comprehensive documentation and examples for easy implementation.

KuzuDB has transitioned to a read-only status, ending active support and prompting users to migrate to FalkorDB, which offers robust support for GraphRAG and agent-centric workloads. The migration process from KuzuDB to FalkorDB involves exporting data with schema discovery to CSV, preserving nodes, edges, properties, and metadata for a seamless transfer. The FalkorDB Rust loader facilitates this by using async batching and streaming CSV, ensuring low latency reads critical for feeding GraphRAG pipelines. The migration is designed to be repeatable, with steps for verifying data integrity and monitoring progress before fully switching operations. FalkorDB is positioned as a reliable alternative due to its focus on low latency graph querying, which is essential for applications involving LLMs and AI-driven knowledge graphs. Pre-migration advice includes treating migrations as repeatable processes, maintaining scripts in version control, and validating data post-migration to ensure consistency. The guide also advises reviewing FalkorDB documentation and aligning migration plans with GraphRAG objectives, with additional resources available for transitioning from other graph databases like Neo4j and RedisGraph.