问HN:有没有人在生产环境中使用OpenAI的Agent SDK?
我正在探索构建一个简单的基于代理的应用程序,并发现了OpenAI的代理SDK: https://openai.github.io/openai-agents-js/。根据我的理解,它封装了很多功能,比如代理循环、函数调用以及与OpenAI MCP服务器的集成,这可能会比使用普通的OpenAI SDK节省我很多工作量。
然而,我有一些疑问:
代理SDK是否过于抽象或难以调试?
有没有人实际在生产应用中使用过它?
为了获得更多的控制和透明度,我是否应该直接在普通的OpenAI SDK上实现逻辑?
感谢任何见解。
查看原文
I'm exploring building a simple agent-based app and came across OpenAI's Agent SDK: https://openai.github.io/openai-agents-js/ . From what I understand, it wraps a lot of functionality — like the agent loop, function calling, and integration with the OpenAI MCP server — which could potentially save me a lot of work compared to using the plain OpenAI SDK.<p>However, I'm wondering:<p>Is the Agent SDK too abstracted or hard to debug?<p>Has anyone actually used it in a real production app yet?<p>Would I be better off just implementing the logic myself on top of the plain OpenAI SDK for more control and transparency?<p>Appreciate any insights.