[Question]: Does Agenta support stream mode?

[coderlxn]

Question

The prompt I'm currently using is very long, and processing it using LLM can take anywhere from several minutes to over ten minutes. If I use a single return value, a timeout occurs. How can I change the way this return result is received to stream mode?

Context

No response

What have you tried?

This setting was not found in the documentation or past issues.

[linear]

AGE-3729

[mmabrouk]

Hi @coderlxn

We do have an experimental streaming feature planned in 96.0

@jp-agenta might be able to add more details there.

One of the challenges we found that multiple provides implement streaming differently. Can you give us more info on:

1. Which language are you using (python/ts/?)
2. How would you expect the streaming to be (any standard in mind)