Descript

  • he Transcript-First AI Audio & Video Editor — Text-Based Cutting, Overdub Voice Clone, Studio Sound, Underlord Co-Editor, and Podcast-to-Social Repurposing in One Doc-Style Workspace1. One-line positioning

    Descript (San Francisco, 2017; Andrew Mason ex-Google; Series C ~$100M, backers a16z/GV) = transcript-first AI audio/video editor​ — you edit by deleting/retyping words in a transcript and the waveform + video cut follow, on top of which sits Underlord (agentic AI co-editor), Studio Sound (one-pass denoise), Overdub (voice clone fix), Eye Contact (gaze correction), Regenerate (retype mis-spoken line → mouth + voice resynced), 30+ lang translate/dub, AI avatar + Veo3.1/Nano Banana video gen. Not Premiere/Final Cut (timeline NLE), not Riverside (remote capture-first), not Opus Clip/Captions (clip-only), not Otter (notes-only). Your 42nd CPT card → ai-productivity-data; pipeline_stage none (spans transcribe → text-edit → AI-clean → clip → dub → export). Table-first, Claude Code skeleton. 2026 trap: AI credits not hours are the real meter​ — Studio Sound ~10 credits/min, Hobbyist 400/mo dies on a 40-min ep, and formerly-“unlimited” features (Overdub regen, eye contact) are now credit-metered.

    2. Core features

    Layer Capability Notes
    Transcribe 25-lang ASR, edit-by-text, filler-word (“um/uh”) one-click remove Delete sentence in transcript = cut in media; 95%+ EN accuracy
    Audio AI Studio Sound (room reverb/HVAC/hard-pops removal), Overdub (clone voice, type fix), Remove Filler Words Studio Sound ≈10 AI credits/min of audio
    Video AI Eye Contact (neural gaze lock), Regenerate (retype → voice+lip resync), Green screen, color, timeline Talking-head only; breaks on head-turn/extreme angle
    Agent Underlord — prompt to “find 10 social clips”, write YT description/chapters, brand layout gallery Credit-metered, agentic co-edit
    Gen media AI avatar (photo/text→Hedra Character-2), AI voices 20+ lang dub, Veo3.1 + Nano Banana image/video gen, royalty-free stock on Creator+ Multi-model router
    Collab Doc-style comments, share link, version restore, multi-seat on Business G2 praise: “like Google Doc for audio”
    Export 720p (Free) / 1080p (Hobbyist) / 4K (Creator+) ; XML to Premiere, FCP No true multicam, no Lumetri-grade control
    Surface 🌐 Web · 🖥️ Win/macOS app · 📱 no real mobile editor · 🔌 Zapier (no public REST API) Desktop app RAM-heavy on 40min+ projects

     

    3. Pricing (2026-08, USD, annual billing; monthly ≈ +40%)

    Tier $/mo (annual) $/mo (monthly) Media hrs AI credits Unlocks
    Free $0 $0 1 hr 100 (signup) 720p watermark-free, limited Underlord, trial AI Speech
    Hobbyist $16​ $24 10 hr 400/mo 1080p, Underlord, Studio Sound, Remove Filler, Overdub regen
    Creator $24​ $35 30 hr (+5 bonus) 800/mo (+500) 4K, full Underlord+20 tools, stock lib, multi-model gen, top-ups
    Business $50​ $65 40 hr 1500/mo Brand Studio, 30+ lang translate/dub, custom avatar, SLA, team roles
    Enterprise Custom Custom Custom Custom SSO/SCIM, data residency (NYT/Spotify tier)

    pricing_type = freemium; starting_price = $0; free_tier = "yes; 1 media hr/mo, 100 AI credits, 720p watermark-free, limited Underlord, trial Overdub — NO 1080p/4K, NO stock lib, NO custom avatar"; pricing_note = "apr-2026-ai-credits-not-hours-real-meter; studio-sound-10-credits-per-audio-min; hobbyist-400-credits-dies-on-40min-ep; formerly-unlimited-overdub-regen-now-metered; creator-24-annual-sweet-spot-solo; business-50-seat-min-for-brand-studio; no-public-rest-api-zapier-only; topups-billed-separately; trustpilot-refund-complaints-jun-2026"

    4. Access Type (v2)

    access_type: web-app, desktop-app

    access_display: 🌐 Web · 🖥️ Windows/macOS · (no mobile editor, no API)

    5. Reviews (split — G2/Capterra high, Trustpilot mixed)

    Platform Score Volume Signal
    G2 4.6/5​ ~850 Transcript editing, Studio Sound, speed
    Capterra 4.7/5​ ~180 Podcast agencies, ease for non-editors
    TrustRadius 4.3/5 ~52 Same shape
    Trustpilot mixed ~2/5​ ~250 AI credit burn, Underlord contradictions, refund promises not honored (Jun 2026 1-star case), app lag on long files
    Aggregate ~4.5/5​ (pro sites) vs ~2/5​ (consumer TP) — Same Ahrefs/Cursor/DeepL pattern: power users love core, billing/credit/support gripes surface on TP

     

    Recurring praise: transcript cut = 4× faster than timeline for spoken-word; Studio Sound salvages bad-room audio; Overdub fixes without re-record; doc-style collab.

    Recurring gripes: credit math opaque​ (Studio Sound 10/min, regen eats credits even on failed Avatar gen), Hobbyist unusable for weekly podcasters, desktop app RAM/scrub-lag on 40min+, Underlord sometimes contradicts itself, support “mostly AI bot”, frequent UI churn, Trustpilot refund-no-show.

    6. Best for / Not for

    Best for Not for
    Podcasters (solo/weekly) editing 20–60min talking-head Cinematic / narrative film (Premiere/DaVinci only)
    YouTube explainer / course / SaaS demo creators Color-grade or VFX-heavy work (no Lumetri/Resolve)
    Content teams repurposing 1 long ep → 10 social clips Pure short-form vertical-only shops (Captions faster)
    Non-editors who fear timelines but can edit docs Tight-budget users: 400 credits evaporate in a day of Studio Sound
    Remote interview pods (record basic + edit in Descript) Need public REST API / BYOK model routing in pipeline
    Agencies doing EN→30-lang dub for clients (Business) Long-form 1h+ with zero human pass — AI clip picks miss joke landing

    7. Competitors (table)

    Tool Lane Starts vs Descript
    Riverside Remote capture-first (4K local rec) $15–21/mo Better source audio, weaker post; “record Riverside / edit Descript” common stack
    Adobe Premiere Pro + Enhance Speech Pro NLE w/ text-based editing (2024+) $22.99/mo Deeper color/comp/multicam, steeper curve, no Overdub/Eye Contact parity
    Opus Clip Long→short clip extractor $15–29/mo Better virality scoring/clip picker, no full edit/transcribe
    Captions Vertical social video + burned captions $10–20/mo TikTok/Reels-only, no podcast timeline, no 4K long-form
    Adobe Podcast Enhance Free bad-room audio cleanup $0 Audio only, no editing/transcript
    Auphonic Loudness/leveling $11/mo Audio only, set-and-forget leveling beats Descript there
    Otter.ai Meeting notes/transcribe $0–10/mo Notes not editor, no video/Overdub
    Cleanvoice Filler/mouth-sound remover $11/mo Single-job, cheaper than Descript if only filler removal

     

Do Not Sell or Share My Personal Information Cookie Settings