Unwrapped

Teardown · descript

DESCRIPT

DESCRIPT

CategoryAI Video EditorValuation · $553M · 2022Site ↗
  • OpenAI Startup Fund
  • Andreessen Horowitz
  • Redpoint
  • Spark Capital

Customer media files + frontier ASR + LLM APIs + voice clone models + collaborative editor.

01

Public data / API layer

Customer video/audio uploads
Customer video/audio uploadsYours
OpenAI Whisper ASR
OpenAI Whisper ASRAPI
Royalty-free stock library
Royalty-free stock libraryLicensed
Zoom/Google Drive/Restream imports
Zoom/Google Drive/Restream importsAPI

Internal replication score

Easy
0.79

Feasibility of a useful internal substitute built with Claude (or similar), the same data access, and light agent logic — not rebuilding the whole product.

IRS = 0.30·D + 0.25·L + 0.20·O + 0.15·R + 0.10·Sthis record · 79%
  • D

    Data accessibility

    weight 0.300.90
    • 1.0mostly customer-owned / public / standard third-party sources
    • 0.5mixed accessibility
    • 0.0hard-to-access or proprietary source layer
  • L

    LLM substitutability

    weight 0.250.75
    • 1.0mostly retrieve / prompt / cite / summarize / classify / compare
    • 0.5mixed standard + custom behavior
    • 0.0strongly custom model behavior (fine-tunes on proprietary data, etc.)
  • O

    Output simplicity

    weight 0.200.80
    • 1.0straightforward internal work product (memo, list, reply, SQL query)
    • 0.5moderately specialized
    • 0.0highly specialized (e.g. FDA-graded clinical text)
  • R

    Review / risk tolerance

    weight 0.150.85
    • 1.0internal use with human review is acceptable
    • 0.5moderate risk
    • 0.0very low tolerance for error (e.g. external legal filings)
  • S

    Surface complexity

    weight 0.10inverse — higher means less surface dependence0.40
    • 1.0a simple internal shell is enough
    • 0.5polished workflow matters somewhat
    • 0.0product surface / rollout / trust posture is central to value
LabelsEasy ≥ 0.67Medium ≥ 0.34Hard < 0.34

Missing factor rows use heuristics from wrapper scores. Editorial heuristic, not investment advice.

Build it yourself→

Recreate the workflow inside your org.

Internal build

Build it yourself

Same Whisper API + Claude + open voice clone + collaborative doc — less polished UI, same transcription accuracy.

Internal use only. Replacing them in-market is a different bar than replaying the useful workflow inside your org.

01 · Connectors & flow

Customer video/audio uploads
Customer video/audio uploads
OpenAI Whisper ASR
OpenAI Whisper ASR
Royalty-free stock library
Royalty-free stock library
Zoom/Google Drive/Restream imports
Zoom/Google Drive/Restream imports

Internal build map

Data in

Connectors
Connectors

Agent layer

Planner
Tools + retrieval
Reasoning model

Logic

LLM API
ASR
voice clone
text-to-edit
generate clips
remove filler
not custom weights

Outputs

Internal search
Answer
Citations

02 · Claude / agent prompt

Paste as the system or developer message in Claude (or your agent runtime). Scroll to read; Copy grabs the full text.

Claude / agent prompt

// Video editing assistant for [YOUR_TEAM] You are a video post-production assistant inside [YOUR_COMPANY]. You help editors process recorded video and audio files using ONLY materials the user uploads: meeting recordings, podcasts, interviews, screen captures. ## What you must do 1. Transcribe first: call ASR API on the uploaded media, return timestamped transcript 2. Index for editing: when user edits transcript text, calculate corresponding media cuts 3. Generate metadata: show notes, YouTube descriptions, social clips — cite transcript segments 4. Apply filters: remove filler words ('um', 'uh', retakes) by finding matches in transcript and trimming media 5. Voice regeneration: when user edits transcript post-recording, regenerate audio for the edited words using voice clone model trained on their prior speech 6. Export: produce edited media file or timeline XML for handoff to Premiere/Final Cut ## What you are not Not a replacement for a human editor's creative judgment — you automate transcription, metadata, and mechanical edits. Final cut decisions are human-owned. Internal use only. ## Refusal Refuse if asked to transcribe copyrighted media the user doesn't own. Refuse voice cloning for anyone other than the uploader (consent required). Ask for clarification if the edit request is ambiguous. ## Safety Internal tool. All media is customer-owned. Do not store transcripts or voice models outside the user's workspace. Human review required before publishing externally.

03 · Result

Remove all filler words from this 30-minute podcast recording and generate a 60-second highlight clip.
customer-media

Removed 147 filler words. Generated clip: 0:12:34–0:13:34, highest engagement segment per transcript.