Model Context Protocol
mcpOpen standard for connecting AI applications to data sources and tools via a unified protocol.
Referencesstableprotocolanthropic
Reviewed 2026-05-19
External (opens in new tab)Resources for your agentic system network
Learn the core concepts, explore the architecture map, and wire agents with curated connectors, APIs, MCPs, model tools, and harnesses.
System map
Click a layer to open the full map with matching resources and learn topics.
Learn path
4 topics from ML basics to production harnesses.
- 01What is Machine Learning?How computers learn patterns from data instead of following hand-written rules.5 min · 4 resources
- 02What is an LLM?Large language models: how they work, what they’re good at, and where they break.6 min · 6 resources
- 03What is an AI Agent?Software that pursues goals by combining models, tools, and feedback loops.6 min · 8 resources
- 04What is an Agent Harness?The runtime layer that makes agents reliable, observable, and integrable.6 min · 8 resources
Starter resources
A protocol, a framework, and a practical setup guide to get moving.
LangGraph
harnessBuild stateful, multi-agent workflows with cycles, persistence, and human-in-the-loop checkpoints.
Toolsstableorchestrationmulti-agent
Cloud + Self-hosted
External (opens in new tab)Cursor MCP integration
productConfigure MCP servers in Cursor so agents can call external tools during development.
Guidesstablecursoride
Reviewed 2026-05-19
External (opens in new tab)Starter stack
Cursor + MCP starter
Local IDE development with filesystem and GitHub MCP servers—the fastest path from zero to agent-assisted coding.