The secret pillar in our hearts (and now, in MUNI)

Descript recently updated its Underlord AI assistant to include a dedicated translation tool, targeting creators who need to localize video content for global audiences. This update simplifies the process of converting a single video into multiple languages by handling both the audio dubbing and the captioning within the same interface. By automating the transcription-to-translation pipeline, the platform aims to reduce the friction of reaching non-English speaking markets.

What's new

The update introduces a translation workflow that converts the original transcript into one of over 30 supported languages. Once the text is translated, Underlord generates a dubbed audio track using a voice that attempts to match the cadence and tone of the original speaker. This process happens simultaneously with the creation of translated captions, ensuring the visual and auditory elements of the video remain synchronized.

Key features of this update include:

• Automatic translation of existing transcripts into dozens of languages.
• AI-generated dubbing that mimics the original speaker’s vocal characteristics.
• Burned-in or sidecar caption generation in the target language.
• The ability to toggle between different language versions within a single project file (see the provider's announcement).

How it fits your workflow

For documentary filmmakers, YouTubers, and corporate editors, this tool replaces a fragmented workflow that typically requires hiring external translators or using separate AI voice cloning services like ElevenLabs. Instead of exporting a script, translating it, and re-importing audio, editors can now manage the entire localization process inside Descript. This is particularly useful for creators who produce educational or interview-driven content where the clarity of the spoken word is the primary focus.

While high-end commercial projects may still require professional human dubbing for perfect nuance, this feature provides a viable solution for mid-tier productions and social media clips. It augments the standard editing process by allowing a creator to test how their content performs in different regions without a significant investment in studio time. The preservation of the original speaker's tone helps maintain brand consistency, making the AI-generated voice feel less like a generic robotic overlay and more like a natural extension of the creator.

What it costs / how to try it

Translation and dubbing are available through the Underlord interface for Descript subscribers. Access to these features depends on your current subscription tier and available AI credits. You can find more details on the specific language support and credit usage on the Descript website.