Skip to main content

LLM (conditional) conversation history awareness

Feature Request

This could be a bad idea but how about LLM memory/history? As in you can chain requests without having to specify every time. Pressing enter would work as usual, but something like option enter / control enter / etc would let you chain requests in the same 'chat'. For example, File A and File B exist Selecting File A, requesting "make into MP4, 1080p" followed by selecting file B, pressing enter, and requesting "this too" will result in a "No request provided" However, Selecting File A, requesting "make into MP4, 1080p" followed by selecting file B, pressing dedicated hotkey (or a continue button next to the return button on the app), and requesting "this too" will result in file B also getting turned to a 1080p MP4 video. Not sure how this will work with online models and API keys, but for local ones, while the model does get unloaded like how it works rn, request history will shortly be kept on cache until you press enter(which resets to default) An additional option which defaults to keeping conversational data (unless a dedicated hotkey is pressed that resets to default) could be interesting.

Posted by lionheo9 ·
Status: Completed1 comment

Log in to comment and vote

Comments1

  • joethephish

    Team•

    May 7, 2025

    Okay, so I have two answers to this:

    1. I have just implemented a “follow up” mode that allows you to reply to the LLM and ask for follow-up tweaks to your original prompt, or respond to requests for clarification. See this feature request which has now been implemented: https://substage.featurebase.app/p/must-have-follow-up-on-prompt

    2. However, most of what you’re talking about can be solved better with the command history, which has been substantially improved in the latest release, since it saves the actual command to be run, allowing you to run it immediately again on another file. Press up on your keyboard to go to the previous command, and press enter with a different file selected to run it again. In the current version this will actually ask AI to re-generate it from the same request, but in the new version it will immediately re-run the same generated command on a new file, allowing you to very quickly perform the same action in multiple files.

    Both of these are now in the beta release (look on our Discord if you’d like to try the beta immediately), and will be rolling out to every user in the next few days.