Devtoolwatch BoardOpenAI Agents SDK review
OpenAI Agents SDK is a Python framework from OpenAI for building multi-agent workflows with tools, handoffs, guardrails, sessions and tracing.
TL;DR
OpenAI Agents SDK is a Python framework for building multi-agent workflows. It is for developers who want a small set of primitives: agents, handoffs and guardrails. It supports the OpenAI Responses and Chat Completions APIs plus 100+ other LLMs, and adds sandbox, realtime and voice agents.
- 30.4kGitHub stars
- +172new stars, week of Oct 25
- ~6.9Mdownloads a week, +15% vs 4-week avg
- #11in Agent frameworks this week
OpenAI Agents SDK at a glance
- Category
- Agent frameworks
- Maintained by
- OpenAI (organization, @openai on GitHub)
- License
- MIT
- Language
- Python, Makefile, PowerShell
- Repository
- github.com/openai/openai-agents-python
- Website
- openai.github.io/openai-agents-python
- npm package
- @openai/agents 0.21.0
- PyPI package
- openai-agents 0.24.0
- Latest release
- @openai/agents 0.21.0 on npm, Oct 20, 2026
- First published
- Mar 11, 2025
What OpenAI Agents SDK does
- OpenAI Agents SDK configures agents as LLMs with instructions, tools, guardrails and handoffs.
- OpenAI Agents SDK lets agents delegate tasks to other agents through handoffs or by using agents as tools.
- OpenAI Agents SDK gives agents tools including functions, MCP servers and hosted tools.
- OpenAI Agents SDK runs sandbox agents that inspect files, run commands, apply patches and keep workspace state across longer tasks.
- OpenAI Agents SDK builds realtime voice agents over WebSocket and voice pipelines that combine speech-to-text, an agent workflow and text-to-speech.
- OpenAI Agents SDK manages conversation history automatically across runs with sessions, including SQLAlchemy, SQLite, Redis, MongoDB, Dapr and encrypted backends.
- OpenAI Agents SDK includes built-in tracing to view, debug and optimize agent runs, plus configurable input and output guardrails.
How it works
You install the openai-agents package in a Python 3.10+ environment and define agents in code. A Runner executes text agents, while RealtimeRunner and VoicePipeline handle realtime and voice agents. Models connect through the OpenAI Responses or Chat Completions APIs, or through third-party adapters such as LiteLLM and any-llm. Sandbox agents run work in a Unix-local, Docker or hosted sandbox client.
How to install and start using OpenAI Agents SDK
With pip
pip install openai-agentsWith uv
uv add openai-agentsWith voice support (pip)
pip install 'openai-agents[voice]'Requirements
- Python 3.10 or newer.
- An OPENAI_API_KEY environment variable for the README examples.
- macOS or Linux for UnixLocalSandboxClient. On Windows, use DockerSandboxClient with the openai-agents[docker] extra or a hosted sandbox client.
First steps
- Create a Python 3.10 or newer virtual environment and install the openai-agents package.
- Set the OPENAI_API_KEY environment variable before running the examples.
- Define an Agent with a name and instructions, then call Runner.run_sync with a prompt.
- Print result.final_output to see the agent's answer, then explore the examples directory.
Who OpenAI Agents SDK is for
- Python developers building multi-agent workflows with delegation between specialized agents.
- Teams that need agents to work in a sandboxed workspace over long tasks.
- Developers building voice or low-latency realtime multimodal agents.
- Developers who want built-in tracing and guardrails without many abstractions.
Limitations to know
- OpenAI Agents SDK in this repository is Python only. A separate JavaScript/TypeScript SDK exists.
- OpenAI Agents SDK's local Unix sandbox client supports only macOS and Linux.
- OpenAI Agents SDK accepts pull requests only from repository collaborators. Others can file issues.
- OpenAI Agents SDK voice and Redis session support require optional extras.
Is OpenAI Agents SDK growing? Usage data
In the week of Oct 25–31, 2026, OpenAI Agents SDK ranked #106 on the Devtoolwatch Board and #11 in Agent frameworks, with a momentum score of 62 out of 100. Its main package had about 6.9M downloads that week, +15% against its own 4-week average.
Show the weekly numbers
| Week of | New stars |
|---|---|
| Aug 9, 2026 | 178 |
| Aug 16, 2026 | 221 |
| Aug 23, 2026 | 207 |
| Aug 30, 2026 | 163 |
| Sep 6, 2026 | 194 |
| Sep 13, 2026 | 178 |
| Sep 20, 2026 | 155 |
| Sep 27, 2026 | 127 |
| Oct 4, 2026 | 152 |
| Oct 11, 2026 | 163 |
| Oct 18, 2026 | 172 |
| Oct 25, 2026 | 172 |
On Hacker News
- OpenAI adds MCP support to Agents SDK, 807 points and 267 comments (Mar 26, 2025)
- Euphony: OSS tool for visualizing chat data and Codex session logs, 7 points and 0 comments (Apr 21, 2026)
- OpenAI rejects 1,200-line community PR for Google's A2A agent protocol, 6 points and 3 comments (Oct 30, 2025)
- OpenAI Agents SDK, 6 points and 1 comments (Mar 11, 2025)
The Board ranks tools by four signals each week: new stars, new stars for the repo's size, added downloads and download growth beyond the market. How the ranking works.
OpenAI Agents SDK alternatives
Other AI agent frameworks on the Board in the week of Oct 25–31, 2026, by momentum:
| Rank | Tool | Stars | Downloads/wk |
|---|---|---|---|
| #1 | Harness²Harness² is a Python framework from Google Research for recursive agent harnessing, refining an agent's prompts, tools, memory and… | 21.7k | – |
| #2 | HyperFramesHyperFrames is an open-source framework that turns HTML, CSS, media, and seekable animations into deterministic MP4 videos, built for AI… | 76.6k | 1.5M |
| #3 | AgentScopeAgentScope is a Python agent framework with an SDK for building ReAct agents and an agent service backend with a web UI. | 35k | 51.4k |
| #4 | AG-UIAG-UI is an open, event-based protocol that standardizes how AI agents connect to user-facing frontend applications. | 16.9k | 4.6M |
| #5 | LangGraphLangGraph is a low-level orchestration framework and runtime for building, managing, and deploying long-running, stateful agents in… | 44.2k | 18.5M |
| #6 | BifrostOpen-source, high-performance AI gateway that works as your LLM gateway, MCP gateway and agent gateway. One OpenAI-compatible API for… | 9.3k | – |
Recent OpenAI Agents SDK releases
OpenAI Agents SDK FAQ
Is OpenAI Agents SDK free and open source?
Yes. OpenAI Agents SDK's source code is on GitHub (openai/openai-agents-python) under the MIT license, so you can use it for free.
How do I install OpenAI Agents SDK?
Run pip install openai-agents. The install section above has the other options from the project's README.
Is OpenAI Agents SDK still maintained?
Yes. The openai/openai-agents-python repository was last updated on Oct 21, 2026, and the latest release, @openai/agents 0.21.0 on npm, came out on Oct 20, 2026.
How popular is OpenAI Agents SDK?
OpenAI Agents SDK has 30.4k GitHub stars as of Nov 1, 2026. It gained 172 new stars in the week of Oct 25–31, 2026, against a 12-week average of 174 a week. Its main package had about 6.9M downloads that week, +15% against its own 4-week average. That made it #11 among AI agent frameworks on the Devtoolwatch Board.
OpenAI Agents SDK vs Claude Agent SDK: how do they compare?
OpenAI Agents SDK and Claude Agent SDK are both tracked by Devtoolwatch as AI agent frameworks. OpenAI Agents SDK has 30.4k GitHub stars and Claude Agent SDK has 8.4k. In the week of Oct 25–31, 2026, OpenAI Agents SDK's main package had 6.9M downloads and Claude Agent SDK's had 25.6M. Claude Agent SDK is a Python SDK for building agents that interact programmatically with Claude Code, including its tools, custom tools, and hooks. Claude Agent SDK review.
OpenAI Agents SDK vs Google ADK: how do they compare?
OpenAI Agents SDK and Google ADK are both tracked by Devtoolwatch as AI agent frameworks. OpenAI Agents SDK has 30.4k GitHub stars and Google ADK has 22.1k. In the week of Oct 25–31, 2026, OpenAI Agents SDK's main package had 6.9M downloads and Google ADK's had 2.4M. Google ADK review.
What are the best alternatives to OpenAI Agents SDK?
Other AI agent frameworks on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: Harness² (#1), HyperFrames (#2), AgentScope (#3), AG-UI (#4), LangGraph (#5). The alternatives table above has their numbers.
What is OpenAI Agents SDK?
OpenAI Agents SDK is a Python framework for building multi-agent workflows. It provides agents, handoffs, agents as tools, guardrails, sessions and tracing, plus sandbox, realtime and voice agents. OpenAI describes it as a production-ready upgrade of its earlier Swarm experiment.
Is OpenAI Agents SDK model agnostic?
Yes. OpenAI Agents SDK is provider-agnostic. It supports the OpenAI Responses and Chat Completions APIs as well as 100+ other LLMs, with third-party adapters for LiteLLM and any-llm.
Does OpenAI Agents SDK support MCP?
Yes. OpenAI Agents SDK lets agents use MCP servers as tools. Recent releases added server-wide guardrails for MCP tools, configurable listing page limits and MCP annotations on converted tools.
Is OpenAI Agents SDK production ready?
OpenAI's documentation calls the OpenAI Agents SDK a production-ready upgrade of its previous agent experiment, Swarm.
When was OpenAI Agents SDK released?
The OpenAI Agents SDK repository was created on March 11, 2025. Version 0.24.0 of the Python package was released on October 15, 2026.
Can OpenAI Agents SDK run on Windows?
Yes, with one caveat for sandbox agents. OpenAI Agents SDK's UnixLocalSandboxClient supports only macOS and Linux. On Windows, use DockerSandboxClient with the openai-agents[docker] extra or a hosted sandbox client.
Sources
- openai/openai-agents-python on GitHub
- OpenAI Agents SDK README
- @openai/agents on npm
- openai-agents on PyPI
- OpenAI Agents SDK website
- Release v0.24.0 (Oct 15, 2026)
- Release v0.23.1 (Oct 2, 2026)
- Release v0.23.0 (Oct 2, 2026)
- Hacker News: OpenAI adds MCP support to Agents SDK
- Hacker News: Euphony: OSS tool for visualizing chat data and Codex session logs
- Hacker News: OpenAI rejects 1,200-line community PR for Google's A2A agent protocol
- Devtoolwatch Board, the week of Oct 25–31, 2026
OpenAI Agents SDK README badge
Maintainers can show the tool's weekly Board rank in their README:
[](https://index.devtoolwatch.com/tools/openai-agents/)How this page is made: Devtoolwatch writes each review from the project's README, releases, package metadata and website, with an AI model, then checks every claim against the source it cites and drops what the sources don't support. The numbers come from Devtoolwatch's daily GitHub snapshots, GitHub's stargazer history, and the npm and PyPI download APIs. Spotted an error? Email [email protected].