Building Modular AI Agents with LangGraph, MCP, and
summary
Claim: MCP turns connectors into discoverable, schema-driven services so agents can validate inputs and swap tools without code changes. Demo builds a LangGraph agent that queries Neo4j via an MCP server (bootstrapped with uvx), uses get_neo4j_schema and read_neo4j_cypher (write disabled), and a local recommendation tool; Gemini generates Cypher which is parameterized and validated. It runs against the IMDb sample graph, returns recommendations (e.g., “40 people who loved Mission Impossible”), and the pipeline works end-to-end. allowed tools
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:56Why use knowledge graphs
- 6:09Understanding AI agents
- 7:38Integration chaos and MCP
- 9:25What is MCP
- 12:39LangGraph for orchestration
- 14:03Neo4j MCP server demo
- 16:34Code walkthrough
- 25:50Q&A and closing
related talks
53mAgents, Tools, and Mcp, Oh My! Next Level AI Concepts for Developers
01h 59mBuilding Agentic and Multi-Agent Systems with LangGraph
45mAgents, Tools, and MCP, oh my! Next-level AI concepts for developers
36mPutting AI Agents to Work: A New Era of Open Source Connectivity with MCP
29mBuilding Interoperable Agents: When A2A meets MCP
01h 03mAI Coding Agents and how to code them
44mBeyond the Prompt: How MCP Connects Language Models to Real Tools
01h 03mAgents Everywhere