Too many MCP servers are simply glorified REST wrappers, regurgitating APIs that were designed for SDKs, not agents. This leads to confused LLMs, wasted tokens, and demonstrably poor performance. If you've ever pointed an MCP generator at an OpenAPI spec and called it a day, this talk is your intervention.

Like any product, great MCP servers are the result of careful design. This talk shares the hard-won lessons from creating FastMCP, the most popular framework for building MCP servers (and yes, for generating them, too). The secret is to stop thinking about endpoints and start thinking about products. We will cover the three pillars of agent-native product design—Discovery, Iteration, and Context—providing an actionable framework for curating context into small, highly effective surface areas that lead to better AI outcomes.

Jeremiah Lowin, CEO of Prefect
https://twitter.com/jlowin
https://www.linkedin.com/in/jlowin
https://github.com/jlowin

AI Engineer

MCP servers often fail because developers treat them like REST APIs instead of designing for agent limitations. Agents struggle with discovery (expensive token enumeration), iteration (slow round trips), and context (limited memory). Best practices include: design for outcomes not operations (one tool = one workflow), flatten arguments to primitives, provide clear instructions and examples, respect token budgets (aim for <50 tools), and curate ruthlessly. Avoid converting REST APIs directly to MCP—bootstrap with them but refactor for agent-specific interfaces. Think of MCP servers as user interfaces for agents, not infrastructure.

Your MCP Server is Bad (and you should feel bad) - Jeremiah Lowin, Prefect