M365-Copilot2API is currently undergoing automated testing with real agents. I’m happy to review your project implementation for free.

I finally thought of this


You see, now AI itself can quickly identify and analyze the actual error messages
image

However, the task may not be real. If you have a task or project you’d like me to work on (as a test),
I’ll send it to you once I’m done.
I’ll implement your idea for free, or maintain/review your project.
You only need to tell me:
The project’s goal
The project URL
The delivery method
That’s all!
(It may not turn out perfectly, since it’s all AI-powered.)

2 Likes

Do everything you want to do

1 Like

I have no inspiration right now and don’t know what project to work on.

Give it a try.

The M365-Copilot2API project itself is indeed an API project designed for Microsoft 365 Copilot. Microsoft also currently provides Microsoft 365 Copilot Agent Evaluations, which supports sending prompts to deployed agents, retrieving responses, and performing automated evaluations.

1 Like

This isn’t really my use case.
It’s more about scoring the LLM’s output itself, not testing the gateway.

I’m not quite sure how the mapping models are configured in version 0.7.2. The model names from previous versions are no longer visible. How do I select a model?

Actually, you can just use the current model—set it to xhigh and it’s basically M365’s sol.

where get the account?