Transcript-based video and podcast editor metered by media hours and AI credits, with export resolution and watermark-free export gated by plan.
Dialogue-led editing where the budget has to be visible: every tier publishes its media hours, AI credits, export resolution, upload ceiling and storage, so throughput can be modelled before buying. It suits buyers who want model training under a switch rather than a clause, because the security page states project-information sharing is explicit opt-in and transcription-data sharing is disabled by default and changeable in account settings. It also fits teams cloning voices with a paper trail, since the terms define a Consenting Speaker and prohibit submitting audio that imitates a voice that has not consented.
Anyone publishing from the free plan: free exports are 720p, watermark-free export begins at Hobbyist, and the allowance is one media hour a month with a one-time 100 AI credits. It is wrong for pipelines that need finished files out of the API, because the API page states direct media file exports are not supported yet and calls the API early access. Stock media is a trap for commercial work: the terms say Storyblocks materials allow commercial use while other stock libraries may not be available for commercial use without a licence.
Pros
Cons
One seat, one media hour a month, 720p local and web-link export, one-hour web links, 5GB of storage and a one-time 100 AI credits, with most AI features marked only as Limited. Watermark-free export is not included on Free and starts at Hobbyist. The FAQ states no credit card is required, so this is a standing free plan rather than a timed trial.
The terms say Input and Output may be used by Descript to train, improve and develop its models, with the parenthetical unless you have opted out, and bar third-party providers from training on the same data. The security page holds the controls: sharing project information requires explicit opt-in, transcription-data sharing is disabled by default and can only be enabled by you, and the setting can be changed anytime.
The terms define a Consenting Speaker as someone who has given both the customer and Descript permission to use recordings of their voice to train an AI Voice, and prohibit submitting audio containing or imitating a voice that has not explicitly consented. Likeness is handled only as a licence grant over names, usernames and likenesses, plus a general warranty that you hold all necessary rights to your content.
Local and web-link exports are 720p on Free, 1080p on Hobbyist and 4K on Creator and Business, with web links lasting one hour on the first two tiers and three hours on the others. Media hours run 1, 10, 30 and 40 per month; AI credits 100 one-time then 400, 800 and 1,500; maximum upload size 1GB, 10GB, 20GB and 50GB.
Facts sourced from the vendor’s own site, last verified 2026-10-24.
The Weekly Source
One email a week. New listings, repriced tools, and the ones we would skip - including this one if its pricing changes.
