FastMCP FastMCP
Sign In
Back to Blog

OpenAI AgentKit with MCP: A Revolution in AI Agent Development

• 6 views

On October 6, 2025, at DevDay, OpenAI introduced AgentKit — a comprehensive toolkit for building, deploying, and optimizing AI agents. This isn't just another API update, but a full-fledged platform that changes the approach to creating agentic systems.

What is AgentKit?

AgentKit is a complete set of building blocks for developers and enterprises to build, deploy, and optimize agents. Until now, building agents meant juggling fragmented tools — complex orchestration without versioning, custom connectors, manual evaluation pipelines, and weeks of frontend work before launch.

Key Components

1. Agent Builder

A visual canvas for creating and versioning multi-agent workflows with drag-and-drop nodes, tool connections, and custom guardrail configuration. Sam Altman called it "like Canva for building agents".

What's particularly impressive: OpenAI engineer Christina Huang built a complete AI workflow and two AI agents live on stage in just 8 minutes.

2. ChatKit

A simple embeddable chat interface that developers can use to integrate chat functionality into their apps with branding and customization capabilities.

3. Evals for Agents

Tools for measuring AI agent performance, including:

  • Step-by-step trace grading
  • Datasets for testing individual agent components
  • Automated prompt optimization
  • Ability to run evaluations on external models directly from the OpenAI platform

4. Connector Registry

A central place for administrators to securely connect agents to internal tools and third-party systems through a control panel while maintaining security and control.

MCP Integration

One of the most important aspects of AgentKit is support for Model Context Protocol (MCP), which allows agents to connect to various data sources and tools through a standardized protocol.

Why does this matter? MCP servers provide agents with access to:

  • Databases and internal company systems
  • External APIs and services
  • File systems and storage
  • Specialized tools

On fastmcp.me, there are many ready-made MCP servers available that can be integrated with agents built through AgentKit. This significantly accelerates development and allows agents to access data and functionality without the need to write custom integrations.

Real-World Use Cases

Klarna built a support agent that handles two-thirds of all tickets, while Clay 10x'd growth with a sales agent.

The Ramp team went from a blank canvas to a buyer agent in just a few hours — what previously took months of complex orchestration and custom code.

Competitive Positioning

The launch highlights OpenAI's push to increase developer adoption by making agent building faster and easier. It also signals a competitive move against other AI platforms racing to offer integrated tools for building autonomous agents for enterprises.

AgentKit directly competes with tools such as:

  • Zapier
  • n8n
  • Make
  • LangChain
  • Other no-code/low-code automation platforms

Availability

AgentKit is available in beta today. All core components are integrated into the OpenAI platform and ready to use.

Conclusion

AgentKit is what the AI agent ecosystem has been missing: a full-fledged platform for creating production-ready solutions without having to piece together a puzzle of disparate tools. Combined with MCP servers available on fastmcp.me, developers get a powerful stack for building agents of any complexity.

As Sam Altman said: "This is all the stuff that we wished we had when we were trying to build our first agents".

Article based on the OpenAI DevDay announcement from October 6, 2025