Descript

Kapwing and Premiere edit videos as visual timelines — you cut on the waveform or video track. Descript edits videos as transcript text — delete a word from the transcript and the corresponding video clip disappears, add a typo correction by retyping the transcript and Descript synthesizes the corrected audio using your trained voice model (Overdub). This is fundamentally different for talking-head content (podcasts, interviews, YouTube essays) — Descript turns 'finding the right take' into a find-replace operation. For B-roll-heavy cinematic editing Premiere wins; for spoken content Descript wins.

Descript has a free tier with 1 hour of transcription per month, watermarks on exports, and limited Overdub voice cloning. Pro is $24/month ($16/month annual) for 30 transcription hours, no watermarks, full Overdub and screen recording. Enterprise tiers add multi-user collaboration. Compared to manual editing in Premiere Pro ($22.99/month) Descript is similar in price but radically different in workflow. For occasional podcasters the free tier covers a single show; weekly creators upgrade to Pro.

Descript uses an in-house ASR model with reported 90-95% accuracy on clear-audio English (closer to OpenAI Whisper's quality). Accents, jargon and noisy audio drop accuracy. The transcription editor lets you correct errors and Descript learns word-frequency for your account over time. For multilingual content accuracy is more variable. The Overdub voice cloning requires 10 minutes of clean voice training and produces convincing single-sentence overdubs but is detectable at longer durations.

Descript's editor and rendering are local (it ships native macOS / Windows apps); transcription requires uploading audio to Descript's servers for cloud ASR. The PWA listed on pwa.directory is the web companion / installable launcher. For full offline workflow the native desktop app is the right choice. Once a project is transcribed and downloaded locally you can edit offline; new transcriptions need a connection.

Descript is the dominant podcast editor in the indie podcast community — it is recommended by Joe Rogan-tier studios (for first-pass editing before professional polish in Pro Tools), used by hundreds of indie podcasters publishing weekly, by YouTubers doing essay-style content (CGP Grey, Wendover Productions reference Descript-like workflows), and by content marketers producing repurposed clips. The text-based editing paradigm has been imitated (Adobe Premiere Pro added a Speech-to-Text panel) but Descript remains the clear leader.