From AI Workflows to Graph-Native Reasoning: A Dify Plugin for Querying Neo4j
summary
Open-source Neo4j query plugin for Dify that runs parameterized Cypher inside agentic workflows, authenticates to Aura or self-hosted instances, and enforces a pre-flight write-block (writes disabled by default). Used with Gemini for text-to-Cypher, an entity-linking step (Linnaeus) for ontology IDs, and multi-hop (up to three-hop) graph queries to rank target genes from integrated sources like DrugBank, STRING and UK Biobank. One-click install, ~2,000+ downloads, and deployed to ~100 enterprise users.
AI-written summary · about
chapters (10)
- 0:00Introduction to Dify
- 2:09What is Dify and its features
- 4:15Missing Neo4j integration gap
- 6:01Knowledge graph use case
- 10:18Neo4j plugin features
- 13:27Sample workflow demo
- 15:58Challenges with LLMs and graphs
- 17:46Ontology and entity linking
- 19:51Conclusion and plugin benefits
- 22:22Q&A session
related talks
31mIntegrating Neo4j with Google Gemini: Architectural Patterns for AI Agents
30mAgentic AI for Enterprise Knowledge Graph Management
30mPerformant GraphRAG Improving Neo4j to Drive Autonomous AI ...
22mBuilding Cypher Queries With Open-Source LLMs
08mHarness Graph Power With Agentic AI: An Insights Discovery Use Case
30mLive-Coding Graph-Native Agents: Neo4j AuraDB + Microsoft Agent Framework in Action
30mAgentic RAG Meets Neo4j: Multi-Hop Reasoning Over Scientific Knowledge Graphs
32mLive Q&A