An Application Programming Interface is a defined contract for how two pieces of software communicate. An API specifies what operations are available, what inputs they accept, what outputs they produce, and what errors they can signal. APIs enable software to be composed from independently built components, which is the foundational pattern behind modern internet-scale software.
APIs come in many architectural flavors: REST APIs (the dominant internet style), GraphQL, gRPC, SOAP, WebSocket-based protocols, and increasingly MCP for AI agent integration. The right choice depends on consumer needs, performance requirements, and ecosystem conventions.
For AI coverage specifically, "API" often refers to commercial language model APIs — Claude API, OpenAI API, Gemini API — that expose model capabilities to developers. The pricing and rate- limiting structure of these APIs has become a significant procurement concern for enterprise AI deployments. API costs, not model capabilities, are often the binding constraint for scaling production AI workloads.