Descript

Overview

Descript is a video and audio editor that treats media as a document. Founded in 2017 by Andrew Mason — previously the founder and CEO of Groupon — the San Francisco company transcribes whatever you record and lets you edit the recording by editing the text: delete a sentence from the transcript and the corresponding video and audio disappear from the timeline. For talking-head content, interviews and podcasts, that inversion removes most of the reason people avoid video editing.

The AI layer is now branded Underlord, an in-app co-editor rather than a separate generator. It responds to prompts — trim the dead air, find the best 60-second clip, tighten this section — and proactively suggests cuts, chapters and titles based on the transcript. Alongside it sit the features Descript has been refining for years: automatic filler-word removal, Studio Sound for cleaning up bad microphone audio, AI-generated social clips, and translation and dubbing across 30+ languages on higher tiers.

Descript has raised roughly $100M+ across its rounds, with a $50M Series C led by the OpenAI Startup Fund, and investors including Andreessen Horowitz, Redpoint and Spark Capital.

The honest limits are metering and performance. Every paid plan caps monthly media hours and AI credits, so a team producing daily video will hit ceilings that a flat-rate NLE subscription would not impose — the Business plan’s 40 media hours is not a lot for a real content operation. Long projects with many media files remain the most common source of complaints: lag, playback stutter and occasional sync problems on heavy sessions. And while Underlord is genuinely useful for assembly, Descript’s timeline is not a substitute for Premiere Pro or DaVinci Resolve when you need frame-level control, colour work or complex compositing. Used for what it is best at — getting spoken-word content to publishable quality fast — it has few real rivals.

Key Benefits

• Near-zero learning curve: Anyone who can edit a Google Doc can produce a clean cut on their first attempt, which is why marketing and podcast teams adopt it without hiring an editor.
• Prompt-driven assembly: Underlord turns the tedious first pass — removing ums, finding highlights, adding chapters — into a single instruction.
• Audio rescue: Studio Sound meaningfully improves recordings made on laptop mics or in bad rooms, salvaging content that would otherwise be unusable.
• One tool end to end: Record, transcribe, edit, caption, clip and publish without exporting between applications.
• Multilingual reach: Built-in translation and dubbing let a single recording become a dozen localized versions.

Use Cases

1. Podcast production — Record remotely, remove filler words automatically, clean the audio with Studio Sound and export episode audio plus social video clips.
2. Marketing and demo video — Turn a screen recording and a script into a captioned, branded product walkthrough without a dedicated video editor.
3. Repurposing long-form content — Generate vertical shorts and highlight clips from a webinar or interview for social distribution.
4. Course and internal training material — Produce and update lesson videos by editing the transcript, so re-recording a whole take is rarely necessary.