You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Ability to have the process keep working while you walk away, with an auto prompt like "Keep integrating and improving the app"
Would love to be able to walk away or work on something else while bolt keeps plugging away. Maybe a button that says "Automate" and when you click it, it would ask for the number of turns (EG:10). It would then auto prompt "Keep integrating and improving the app" or whatever prompt the user sets.
The text was updated successfully, but these errors were encountered:
I dont think that makes sense at the moment, because you dont now what the output is. maybe the 3rd iteration is the best and you keep going till 10 and it messes up the whole project. Even if it was perfect at round 3, you got nothing at the end.
Then how about a separate LLM (like 4o checking Sonnet, or vice versa) checking the progress and directing the agent what to do by sending it as a prompt?
Seems like it could be more effective to allow opt-in of “I’m feeling lucky” behavior, with some ability to specify how far it should go/whether it should hit the brakes on the first error. It’s an interesting solution to a coffee break where you’re just trying something out.
Ability to have the process keep working while you walk away, with an auto prompt like "Keep integrating and improving the app"
Would love to be able to walk away or work on something else while bolt keeps plugging away. Maybe a button that says "Automate" and when you click it, it would ask for the number of turns (EG:10). It would then auto prompt "Keep integrating and improving the app" or whatever prompt the user sets.
The text was updated successfully, but these errors were encountered: