Interactive dry-runs/introspection? #719
-
I've used gptel for a bit in org mode and it's great, but I'd like to know exactly what gets sent as the prompt. Right now it seems like the only way to be sure what is sent is to explicitly select a region. In the readme, it states there's support for "introspection, so you can see exactly what will be sent. Inspect and modify queries before sending them". I then found the function |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 4 replies
-
Yes, turn on gptel's expert commands with it's a mature feature but a little under-documented at the moment because the "correct" way to present the UI for this feature also puts the onus on me to educate users on transient levels. If you don't mind, could I ask you a question? Are you aware of transient-levels, a feature available in every transient menu? |
Beta Was this translation helpful? Give feedback.
Yes, turn on gptel's expert commands with
(setq gptel-expert-commands t)
. Then you'll find dry-run options in the transient menu. You can view the payload to be sent as Lisp or JSON, and edit it in place and resume the request if you'd like.it's a mature feature but a little under-documented at the moment because the "correct" way to present the UI for this feature also puts the onus on me to educate users on transient levels. If you don't mind, could I ask you a question? Are you aware of transient-levels, a feature available in every transient menu?