MCP

Engineering Blog

From Discovery to Execution: How AI Agents Can Dynamically Install Tools

General-purpose agents often rely on tools served by Model Context Protocol (MCP) servers to accomplish things for humans. MCP has quickly become the go-to protocol for enabling AI agents to access a wide range of tools. At a high level, a large language model (LLM) acts as the agent’s “brain,” while tools give it the […]
Read more
Engineering Blog

Converting OpenAPI Specs to Agent Tools

Before we dive in, let’s clarify what we mean by “agent tools”. Agent tools are local functions whose descriptions and parameters are exposed to an agent. The agent can then supply all the parameters to perform a “tool call”. Then, the agent framework (CrewAI, AWS Strands, Langchain, etc.) orchestrates running the tool function with those […]
Read more
AI merchant service

Agent Tool Calling Chaos when MCP Server Tools have Optional or Nullable Parameters

When building MCP servers, you quickly realize something: the devil really is in the details. One issue we’ve encountered is how differently LLMs and agent frameworks handle optional and nullable parameters in MCP server tools. At first glance, this might sound like a subtle implementation detail. But as we learned—through multiple iterations across OpenAI, Anthropic, […]
Read more