Notifications and long running operations

[faisalil]

Pre-submission Checklist

• I have checked that this question would not be more appropriate as an issue in a specific repository
• I have searched existing discussions and documentation for answers

Question Category

• Protocol Specification
• SDK Usage
• Server Implementation
• General Implementation
• Documentation
• Other

Your Question

In the docs, there is mention of notifications that the server can send to the clients - I see sample implmentation for this in the everything server. My question is about the use case for such a feature: is that idea that these notification updates (or messages) are being sent to the LLM as messages? as an example, if I start running a process and it output some logs that the LLM wants to use to stop the process, is this the kind of senarios this is meant for?

[jspahrsummers]

At the moment, we don't have any notifications that are intended for consumption by the model itself, but this seems like a good idea! It could unlock more agentic use cases to support something like that.

[hasani114]

@jspahrsummers Are notifications tied to specific conversations/threads/messages? I understand they can be via metadata but by default they have no relationship to a specific "run" of the Agent? I'm trying to think about long running agents, where the user may want to start multiple conversations/threads at the same time and in that case it will be beneficial to have notifications per thread that can be visible on the interface instead of general notifications.