Feature Requests

I find that I am constantly extending the pauses between sentences to get my AI narrator to slow down. A feature to set a custom default timing would be nice (for example, setting the pause after a period to a .5 sec by default). It would be above and beyond to give a variance option, such as, setting the pause after a period to a range of .5 to 1 sec. Then allowing it to randomly assign a pause length in that range to each sentence. This would break up the speed and monotony to better reflect the pauses that might occur in natural speaking.

I was five years old, chilling in the living room, minding my own business. Then, out of nowhere, I get this brilliant idea—what if I stick a bead up my nose? You know, as one does at five. No idea what I was thinking, but hey, curiosity got the best of me. Well, surprise surprise, I couldn’t get it out. So, my grandpa goes into full panic mode and rushes me to the hospital, looking like a knight on a mission. But here’s where it gets good—when the doctors tried to help me, I turned into a human barricade. I’m talking full-on wrestling match. I was not having it. Eventually, they got the bead out (probably with some bribery or Jedi mind tricks), but seriously, folks, the moral of the story is: Do NOT stick things up your nose. It’s not as fun as it sounds—and it might just send your grandpa to the ER too.

"Thank you, Dr. AIM! I’m excited to be here and share my journey." "Yes, my mother started the rabbit farm, and I saw how much effort she put into it. Over time, I developed a passion for it, and I wanted to take it further." "My siblings encouraged me to keep pushing, and that motivated me to turn it into a real business." "One of the biggest challenges was managing resources. Raising rabbits requires patience, knowledge, and the right environment." "Starting Dominix Music alongside the farm was tough, but I love music, so I found a way to balance both by setting a daily schedule." "Many young people think success happens overnight. But it takes consistency. If you’re passionate about something, start small and grow from there." "It’s okay to fail. What matters is learning and moving forward." "For Dominic Rabbit Farm, we’re expanding to supply more fresh meat to customers." "For Dominix Music, I want to train more young pianists and build a music academy in the future." "Thank you, Dr. AIM! It’s been an amazing conversation. I appreciate this platform to share my story!"

I have tried two versions of Overdub: 1 with the classic training model, where I recorded 30 minutes of script; 1 with the new, faster training. The first option is FAR superior. Is there still a way to use the old training method to create a new speaker? We're interested in using Overdub, but this would be a dealbreaker.

Hi Descript team, I’d love the ability to train an AI Speaker directly in French. Currently, training with English text results in my AI Speaker reading French scripts with an English accent. Would it be possible to add support for recording AI Speakers using non-English training texts? This feature would be incredibly valuable for multilingual users. Thank you for considering this!

Currently when you use overdub to replace your audio with an AI voice, it only lip-syncs. I want to be able to have the AI voice emulate the inflection of the recorded voice, and the pronunciation, and even the accent. This should come with controls so you can tell the AI overdub to copy or ignore inflection, copy or ignore pronunciation, copy or ignore accent or add a different style, etc.

Allow us to type words that you would generate audio for based on the voice recording we have (helpful for audio malfunctions in live productions that require post-production edits such as TEDx talks)

Would be nice to use voice-clones from ElevenLabs.

Funny