Building API Endpoints as MCP Tools for AI Agents using CrewAI
We’ve always written APIs for devs, for us humans - clean endpoints, fine-grained, neat documentation. Expecting developers to figure out how to stitch them together. But that doesn’t work anymore.
Today, APIs are being used directly by AI agents to get things done. The agents don't look for endpoints. They look for tools that help them complete tasks.
In this talk, we’ll walk through:
Why traditional REST APIs fall short when consumed by agents
How MCP brings three key ideas together: Context, Tools, and Tasks and how this redefines the behavior of APIs.
What it means to design an API as a tool: self-contained, goal-oriented, and context-ready
The role of governance and observability in managing tool-based APIs at scale
A live demo on converting existing API into MCP Server and Endpoints into Tools, showing how an AI agent can use that tool to complete tasks independently.
Building a tech news summarizer AI agent using Gravitee's MCP Tool Server, MCP Inspector, CrewAI and Ollama with Python
#AI #MCP #Agents #API #Observability
