hanlonsrazor

Those building with OpenAI's Agents SDK, how would you describe your experience with the tech so far? LangGraph has recently been giving me trouble with debugging behind the scenes of their abstractions in some more complex multi-agent workflows, does Agents SDK solve this problem while still giving the tools for more complex workflows?

Are there any existing tools to expose APIs via MCP standards? Looking to expose some of our product APIs for integration but dev time for a new server seems excessive. Found https://www.openmcpify.com/ but it's still under development. Any ideas, or am I going down the wrong rabbit hole?