OpenAI Agents SDK

Successor to OpenAI's experimental Swarm project, the Agents SDK is a lightweight Python framework for building multi-agent workflows around a small set of primitives: agents, handoffs that delegate work between agents, guardrails that validate inputs and outputs, sessions that manage conversation history automatically, and built-in tracing for debugging and evaluation. Although built by OpenAI, it is provider-agnostic, working with the OpenAI Responses and Chat Completions APIs plus more than 100 other LLMs. The SDK installs with pip install openai-agents on Python 3.10 or newer, with optional extras for voice agents (including realtime audio) and Redis-backed session storage, and a separately maintained TypeScript version exists. Human-in-the-loop approval flows and sandboxed agent workspaces are supported for long-running tasks. Code is MIT licensed with roughly 28,000 GitHub stars, and it has quickly become one of the standard choices for teams building production agent pipelines on OpenAI models.