Transcript-first clip intelligence

Find the moments
that feel edited
before you cut.

Paste a long-form podcast. Get a scored editorial shortlist with hooks, timestamps, captions, and export-ready files — in one considered workflow.

Subtitle aware Hook scoring GPT curation CSV & JSON export

New analysis

Analyze a
podcast.

Paste any public YouTube episode URL. The pipeline pulls subtitles, scores hook density, and sends only the strongest chunks to GPT for editorial curation.

Defaults tuned for short-form discovery
Results cached for re-runs
JSON + CSV export included
01
↗
Full YouTube link or youtu.be short link
02
chunks
clips

Processing takes 30s–3min depending on episode length. You'll be redirected to a live job page.

How it works

A pipeline in four movements.

01

Transcript

Pulls subtitles first so the pipeline stays fast, readable, and inexpensive.

02

Chunk & score

Ranks overlapping chunks by hook density, emotional charge, and standalone potential.

03

GPT curation

Sends only the best chunks for editorial selection, titles, hooks, and risk assessment.

04

Export

Returns structured JSON and CSV with ffmpeg commands ready to cut.