Descript is an audio and video editor in which you cut and rearrange recordings by editing their transcript as text.
✓ Pros
- Deleting or moving text in the transcript edits the underlying audio and video
- AI voice clone (with recorded consent) can replace a misspoken word without re-recording
- Screen recording, multitrack editing, captions and publishing in one app
- API and MCP server let paying users automate edits such as filler removal and captions
✗ Cons
- Transcription does not support Japanese, Chinese or other non-Latin-script languages
- Filler-word detection and the transcription glossary work only in English
- Monthly media hours and AI credits cap usage on every plan
- Only one spoken language per file can be transcribed
| Free tier | ✓ Free tier |
| Pricing model | subscription |
| Price (Hobbyist（年払）) | $16 USD |
| Features | |
| Languages | en, fr, de, es |
| API | ✓ Available Docs ↗ |
| Pricing Plans | Free$0/月60分/月、720p書き出しに透かし Hobbyist$16/月年払。月払いは$24。10時間/月、1080p透かしなし Creator$24/月年払。月払いは$35。30時間/月、4K書き出し Business$50/月年払。月払いは$65。40時間/月、AIクレジット1,500/月、最大5人 Enterprise要問合せSSO/SCIM |
| Platforms | |
| Integrations | YouTube, Spotify, Adobe Premiere (XML export), Slack |
| Homepage | https://www.descript.com |
Overview
Descript suits podcasters and solo video creators who think in sentences rather than waveforms: deleting a line in the transcript cuts the clip. An AI voice clone, made only after a recorded consent statement, can retype a misspoken word. The main limit for Japanese users is transcription: about 25 Latin-script languages are supported, Japanese is not, and filler-word detection works only in English. Plans are capped by monthly media hours and AI credits, so heavy editors should check those caps. An API and MCP server, still in early access, allow scripted edits.
