Feature Requests

It seems I can't edit captions when there's an edit break in the timeline. I want to be able to drag and reposition timing of captions and correct/generate new script and captions in places where the transcription was missed. Right now I can't do this if there are edit boundaries on either side of a region.

Crashes happen, that's fine - but if I use Underlord to make a series of edits which I can then go through and commit or revert - usually 100+ in a podcast - and a crash happens, the edits remain as ignored text but I don't seem to have any way of going back to the list of edits to continue accepting them easily. This is super annoying. If there is a 'view uncommitted edits' or 'view all instances of ignored text' toggle I can't find it. If there isn't one there should be.

When manually aligning text to video, adjustments to preceding or following sentences should not alter the existing alignment. This repetitive process is time-consuming and inefficient. There should be an option to enable or disable automatic corrections based on user preference.

Underlord transcribes U.S. as being the end and start of a sentence (U. S.), and treats the commas in numbers as a grammatical punctuation (intead of 2,000 it's 2, 000). Please work on making it able to distinguish those. Thanks!

When I translate audio, I have a native speaker on staff review and correct. Sometimes it necessitates a change to the original audio (i.e. removing a section that isnt relevant). This means I have to regenerate the entire translation, which isn't the same every time, so my editor has to re-review and correct the translation again, leading to an endless loop.

When I translate audio, I have a native speaker on staff review and correct. Sometimes it necessitates a change to the original audio (i.e. removing a section that isnt relevant). This means I have to regenerate the entire translation, which isn't the same every time, so my editor has to re-review and correct the translation again, leading to an endless loop.

When transcribing numbers like "10,000" or "120,000," Descript adds a space after the comma, yielding "10, 000" and "120, 000."

For Court Q. and A. designation between speakers.

I sometimes use Descript for my Hindi-English bilingual audio transcription, in which I need English-spoken words in English and Hindi-spoken words in Hindi. Till a few weeks ago, Descript was giving similar transcriptions like English words in English and Hindi in Hindi. But now I am facing problem in it. Now if I'm selecting the transcription language as ‘Hindi’, then it's not transcribing the English speech in English. It's now transliterating English into Hindi. That's the issue. Could you please fix it as how it was working before? Please try to revert that feature. Thank you.

I would like to add Vietnamese languague as it is not available right now