We also had to reverse engineer the SDKs because we couldn't find any good source of what Streamable HTTP really supposed to do and our big take away from this experiment was that MCP is an experiment.
Why would anyone want to write a non-scalable wrapper around a service that already is well documented using OpenAPI endpoints is beyond me.
Anyway, we ended up implementing it just because but I already know it is a mistake and a potential source of many hours wasted by our engineering team.
Why would anyone want to write a non-scalable wrapper around a service that already is well documented using OpenAPI endpoints is beyond me.
Anyway, we ended up implementing it just because but I already know it is a mistake and a potential source of many hours wasted by our engineering team.