I am reporting this a bug or defect as it seems unintentional.

When sending a prompt to Brain, the user has the option to stop processing the prompt to make corrections, add additional context or just to finish the prompt they mistakenly sent early. This action clears the prompt that was being processed before it is stored to history, forcing the user to fully rewrite the prompt instead of just making a few changes.

Other AI desktop tools allow the user the stop the processing of a new chat without clearing the text and losing the thought.

Please also consider allowing the user to change the send prompt from the enter/return key to something else, like cntl+return or alt+return. This will help users from mistakenly sending prompts early, wasting AI time/credits, and losing the initial prompt.