Feature Requests

I wanted give my feedback regarding the Ai credits. It's not fair compared to the amount of minutes I have for the same service plan.

I'm a brand new user trying to edit podcast files. I allowed underlord to help me - it froze the project, combined two user videos and completely "lost" - one of them (who now has no face and no audio). All the editing I'd done is gone. And now I'm getting messages that say I've used half of my upload time on the "hobbyist" plan and urging me to upgrade. And also it says I've run out of AI time and I should buy more. I paid for a year in advance 2 days ago. This is just a DISASTER so far. I'm willing to try again and not ask for my money back - but is there a way to "re-set" the account from scratch and I can just ignore the underlord?

when use video regenerate . the regenerated video is not align with the adjusted size. it is still the orignal size. I scale my video. the regenate one is always smailer , troubled me to adjusted it 30+ times. When I use the video regenerate feature, the regenerated video does not retain my previous scaling or size adjustments. It always reverts to the original size, so I have to manually resize it every time. This is very frustrating, especially when I have to do it 30+ times. Please improve this so that regenerated videos keep the adjusted size and position.

AI features like 'removing retakes' is one of the most useful features - but the AI used often makes a lot of mistakes, leaving in retakes or removing things that aren't retakes. I would like to be able to pick which model is used to find the one that works best for me.

Inadequate Edit for Clarity

I have a client that I edit for that uses two cameras to film herself. So in my editing I jump between the camera angles whenever I cut the timeline. Because of how she prefers the edits to look, the shorten word gaps feature comes in pretty handy to do a rough cut up of the timeline that I can then refine, but it doesn't add scene markers (should be a toggle because come on guys why doesn't that already exist). So I've tried to use Underlord to mass add those in for me to make my life a bit easier. I've done this across three projects and it has not worked properly every time. The ask is simple: "Please apply scene markers to every cut in the timeline" This timeline has about 250+ cuts in it, but I never even get 100 scene markers placed. So I still have to manually add in most of them. Then to add insult to injury? Underlord starts applying smart transitions to everything. I've even tried specifying in the ask to not include any transitions, and it does this anyway. And yes, I have smart transitions on new scenes disabled. When I ask Underlord to remove them, it says its unable to and tells me it's added by default (despite me having that setting disabled already). So now I have to go through all of the scenes it did make and manually remove the smart transitions, and also still make most of the scenes as I do. What is supposed to be saving me time on the tedious parts of editing is now giving me more tedious work and wasting my time. As for an official request to come out of this, there are two: Have an option for the word gap shortening feature to place a scene marker where it makes cuts. For anyone who has multiple camera angles in use this seems like a no brainer add and would be incredibly useful. My god please make sure if underlord can edit the transcript it's actually doing what you ask, and if it can't do that don't just do what it thinks is close. If I can add just a scene marker, Underlord should be able to add just a scene marker. If I ask it not to apply transitions, there is no world where it should be mass applying transitions. If it does apply transitions, it should be able to remove them if asked to do so. I understand the feature is still a beta, but this is a glaring QA issue that needs to be addressed.

Laughter in a recording is not transcribed, and shows up as "silence." Thus bulk remove of silence removes laughter, and other non-word communication (gasps, sighs, etc.). Ideally transcription would tag common exclamations as Laughter, Sigh, etc. Barring that, it would be nice if it could distinguish between actual silence, and non-word sounds in transcription so that shorten silences wouldn't cut out things like that, or could be at least be configured not to. It seems like the "tagging" functionality would be a good match for this, however that does not work on Multitrack, and this feature would be needed in Multitrack as well.

Spontaneously recorded shows often include laughter, which which sounds creepy and unnatural after Studio Sound, and which edits poorly with Shorten Gaps.

When using Descript on a browser, it would be great to be able to access the chat conversation with Underlord even after closing the browser. There are times when I need to shut down the browser and/or computer to keep my system running smoothly, but unfortunately, that means I lose the conversation I've had with Underlord and all the work we've created together. Is there a way to autosave a chat conversation periodically in order to access it in the future?

Support option to use user defined LLM resources. Many enterprises already have LLM's contextualized for their environment using RAG and other content guardrails so that when a prompt is used to create content that the content is not generalized, but is instead already optimized to create content for that enterprise ecosystem. This would speed up ability to create scripts that can be used with auto-generating tools for visuals. Perhaps this needs to implemented using MCP or similar given rise of agentic AI flows.