Live-Coding Graph-Native Agents: Neo4j AuraDB + Microsoft Agent Framework in Action
summary
Connects Microsoft Agent Framework with Neo4j AuraDB and Azure AI Foundry to give agents a true GraphRAG-style graph memory, ingesting three real contracts via LLM extraction to JSON, creating nodes/relationships and generating embeddings per excerpt for semantic search. Uses Azure OpenAI (with easy swap to Anthropic), a text-to-Cypher retriever that turns natural language into Cypher queries, shows runnable code, live demo of queries and monitoring via OpenTelemetry — pipeline works end to end.
AI-written summary · about
chapters (6)
- 0:00Introduction and speaker intro
- 3:34Why Microsoft Agent Framework
- 9:41Building first agent demo
- 15:45Agent workflow basics
- 24:41Graph-native agent with Neo4j
- 29:27Demo wrap-up and resources
related talks
58mAdvanced Autonomous AI Agents: Memory, Collaboration and Evolution in Action
31mAura Agents in Action
26mAI powered automation & multi-agent orchestration
46mAI Agents for the Rest of Us
12mLeveraging Neo4j Graph Data and AI Across the App Development and Deployment Lifecycle
29mThe AI Agent Memory Landscape
01h 54mBuilding Code First AI Agents with Azure AI Agent Service
14mUsing Graphs to Fine-Tune Agents