Apache HugeGraph is a fully open source graph database stack: the database, distributed mode, analytics, web UI, and the HugeGraph-AI toolkit, all under one Apache license and one community. In this episode, contributor Himanshu Verma explains why graphs matter for AI: ordinary retrieval finds text that looks similar, but some questions are about how things are connected, and a graph lets the model look up real facts instead of guessing. We talk about how HugeGraph compares to Neo4j, JanusGraph, and Apache AGE, its support for both Gremlin (Apache TinkerPop) and Cypher, and the 1.8 release work, including the MCP server, GraphRAG, text2gremlin, Helm charts for Kubernetes, and the new Hubble UI. Himanshu also walks through a demo: you upload your documents, an LLM pulls out the people, companies and links between them for you to review, and then you ask questions in plain language. It can run on a local model through Ollama, so private documents stay on your machine. We finish with how to get involved, from filing issues and fixing docs to taking a good first issue.
Learn more: https://hugegraph.apache.org
Prefer video, that’s on our YouTube channel.