Explain the problem as you see it
Currently there's no indication of how far along the user's AI calls are.
Why is this a problem for you?
In some cases it's not clear whether the call has 'hung'/failed, while the spinner keeps spinning.
Having a better sense of the API call progress would help me decide whether to divert my attention to something else temporarily, or wait.
Suggest a solution
One improvement could be to have two distinct 'progress' states, reflected in the spinner animation.
- Prompt is being sent to GPT
- Tana is receiving GPT output.
One way of differentiating between states could be to have the node icon pulse (on/off) while data is being sent, and spin while data is being received.