Agent Platforms

Agent platforms are managed, enterprise-grade environments for building, deploying, and operating AI agents at scale. Unlike open-source frameworks that run in your own infrastructure, platforms provide hosted runtimes, built-in observability, governance controls, and marketplace integrations out of the box.

Neo4j connects to these platforms through MCP server integrations, REST APIs, and native driver connections. The knowledge graph becomes available as a tool within each platform’s agent runtime — providing structured relationship data, real-time graph traversal, and vector search that complement the platform’s own data sources.

The Neo4j MCP server is the primary integration point across all platforms, exposing Cypher execution and schema inspection as standard tools. Platforms without native MCP support can connect through REST bridges or platform-specific extension mechanisms.

Agent Platforms

  • AWS AgentCore — Amazon Bedrock-based agent runtime with MCP support, IAM authentication, and AgentCore Gateway

  • Azure AI Foundry — Microsoft’s enterprise agent platform with Copilot Studio and Azure AI Agent Service

  • Databricks Agent Bricks — Databricks-native agent framework with Unity Catalog tool governance and MCP integration

  • Google Gemini Enterprise — Google’s enterprise AI platform with Vertex AI Extensions and Agent-to-Agent protocol support

  • Salesforce Agentforce — Salesforce’s Atlas Reasoning Engine with External Service Actions and native MCP client

GraphAcademy Courses

If you want to learn how LLMs and Knowledge Graphs combine to improve GenAI applications, check out the Neo4j & LLM courses on GraphAcademy.

llm fundamentals