-
Notifications
You must be signed in to change notification settings - Fork 34.5k
Open
Description
We get a lot of user feedback about missing transparency around tokens used (and no transparency around rate limits).
As a step 1, I think we should introduce some UI affordance that tells the user how many tokens they are about to consume (have consumed so far in the session).
LM Studio has nice affordance that shows token use after every message, and the text input also shows how much tokens the current input consumes. Below the input is a simple tokens N / MAX
Zed simply shows an affordance in the title area, which is very minimalistic but I believe gets the job done.
nimoalr, DovieW, jonathonlacher, Frulfump, DanTup and 2 more