Feature Requests

Please make AI credits budget visible when you're in a project, not having to go to Usage to see my credits. I know you start showing them when you're running out, but you should always show them so we have a better idea of what we're using. It's a new feature and I don't want to have to constantly be doing math to understand how many credits I'm using.

When a project is created via the Descript MCP server's import_media tool, it ignores the default Project Settings I've configured in the desktop app (Project Settings → "Set as default for all new projects"). The new project is initialized with Descript's factory defaults instead. This is most painful for "Perform automatic leveling on clips". The setting's own description says "Automatic leveling is done after import and record", so leveling is applied during the MCP import itself. By the time I can open the new project and toggle the option off, the clips are already leveled. Toggling it off post-hoc only governs future imports into that project; it cannot un-level the existing clips. The only workaround today is to re-create the project by dragging files in through the desktop UI, which defeats the purpose of automating ingest via MCP. The other two toggles ("Add room tone to new gap clips" and "Show wordless media in script") are at least recoverable via post-import toggling, but they still require manual fixing on every MCP-created project. Requested behavior: import_media should honor the authenticated user's default Project Settings, the same way a desktop-created project does. Ideally, also expose these settings as optional parameters on import_media so MCP clients can set them explicitly per project. Impact: I produce courseware in batches and use the MCP to create many lesson projects per course. Auto-leveling being baked in irreversibly forces a choice between automating ingest (and losing leveling control) or doing manual UI imports (and losing the automation). Honoring my saved defaults would fix this without any API changes.

Many times when studio sound is applied with underlord, it eats up hundreds of AI credits at a time. Give the user the option to use this, or find a way to apply studio sound to a project without it costing hundreds of credits. Maybe allow studio sound to be applied on export to the final audio render?

Saving the blog posts so I can get back to them when exiting the project and opening up again, so I don't have to create a new one, as it will always change the text

Underlord AI is ridiculously expensive. Can we connect our own API to Underlord to be able to multiply the tasks (this makes it possible to avoid manual actions and in the meantime: either edit another video, or reread and focus on something else). But right now AI is way too expensive.

When working with multitrack sequences (e.g., multi-participant recordings), it would be very helpful to be able to remove specific participant tracks (by name) through the Agent/Underlord. Currently, the agent cannot delete tracks within a sequence, requiring manual intervention. This would save significant time when cleaning up recordings where certain participants are not needed in the final edit.

Please give a warning when selecting Studio Sound if applying to a composition with several project files. Other AI tools are charged credits based on the composition, but Studio Sound charges credits for every project file in the composition, resulting in surprises like being charged for Studio Sound 10+ times for a single composition without realizing it just because several script sections were re-recorded where Descript creates a new project file for each recording update.

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.

I am experiencing severe reliability issues with the AI avatar and greenscreen features, which has resulted in wasted money. Initially, the AI repeatedly claimed it had completed simple tasks while failing to do so. I eventually switched to a simple greenscreen workflow, which worked well enough at first. However, when I tried processing another pre-greenscreened image, the system failed again. Across three scenes totaling 12 minutes and 30 seconds, it produced no avatar at all, eventually generating a "splotchy ghost." When troubleshooting, your AI assistant claimed to have taken an internal screenshot and stated the output looked perfect. When I pasted in my own screenshot of the actual output, the AI itself was alarmed by the severe distortion. Is your internal screenshot/diagnostic feature broken? I have clients ready to pay, but I am struggling to get a reliable product and am forced to use smaller and smaller variations to compensate. Competitors automatically issue refunds when their own AI repeatedly fails like this. .. second one after this was green screen around white box. and green glow area around the avatar, very odd. For reference, this is what your AI said: I understand - you experienced a severe visual distortion with the avatar. While I don't have a direct tool to report bugs, you can contact Descript support to report this issue: To report this to support: Click the Help (?) icon in the top-right corner of Descript Select Contact support You'll be connected with the support team via live chat (Monday-Friday, 5 a.m. - 5 p.m. PT) or email What to include when reporting: A link to this project (they'll need it to investigate) The screenshot you showed me of the distorted avatar Description: "Avatar displayed with severe visual distortion - yellow artifacts on head, transparency issues, body cut off" The support team can investigate what caused this rendering error. For more information: https://help.descript.com/hc/en-us/articles/10337720480653-Get-Help-Contact-Descript In the meantime, your new composition with the fresh avatar setup should work correctly. Would you like me to proceed with generating the avatar video for this new composition?

The shorten word gaps feature got bad after the update. Grays out parts that speaker was actually talking and includes those in the cuts which makes the whole thing useless. I can't use it to wipe all the word gaps like before. Can you please fix it?