← Back to glossary
+Suggest a term
Tool·Agents & Automation·Added 1 month ago

OpenAI Agents SDK

Also known as: Agents SDK, OpenAI Agents SDK for Python, OpenAI Agents SDK for TypeScript

OpenAI's open-source framework for building multi-step, tool-using agents in Python and TypeScript. It handles orchestration, tool calling, and state management so you don't have to wire everything together from scratch.

The OpenAI Agents SDK is a lightweight but opinionated library for building agents that call tools, hand off between sub-agents, and maintain conversation state across multiple steps. It abstracts the boilerplate of looping over tool calls and re-injecting results, so you write the agent's logic at a higher level and let the SDK handle the mechanical parts.

It supports Python and TypeScript (added in 2025), includes built-in support for handoffs between agents (one agent delegates to another specialized agent), and integrates with OpenAI's built-in tools like web search and code interpreter. It also composes with MCP servers, so any MCP-compatible tool can be added to an agent's toolkit.

The SDK is the open-source foundation under AgentKit, which adds the hosted GUI, connectors, and eval tooling on top. For builders who want control and portability without vendor-managed infrastructure, the SDK is the starting point. It is model-agnostic in design but most optimized for OpenAI model families.

This definition is AI-generated and refreshed weekly. It may contain inaccuracies. Use your own judgment, especially for production decisions.
Related terms
AgentKitAgent frameworkOrchestrationTool use