Long-form in. Ranked shorts out.

Fifteen clips are hiding in your last recording.

Paste a YouTube link or upload the file. SnipHawk transcribes it word by word, scores every moment in it, and hands back a ranked list of clips — already reframed, already captioned, ready to export.

Access is opened in small batches. No card, and no obligation.

A SnipHawk project showing twelve ranked clips, a 9:16 preview with a burned-in caption, and export settings for layout, captions, quality and format.
Ninety-two minutes in. Up to fifteen clips out. The lit window on the timeline is the moment it picked; the list on the left is what comes back, with the export happening on the right.

Nobody wants to watch their own podcast twice.

Finding the clips in a ninety-minute recording is a manual job. Watch it back, mark the times, cut each one, reframe it, caption it, export it. Five clips is an afternoon. Most weeks that afternoon doesn't happen, and the episode goes out once instead of nine times.

Every clip costs a full re-watch

The fortieth clip costs the same as the first: another pass over the whole file to find the moment you half-remember.

A transcript isn't a shortlist

A wall of timestamps tells you what was said. You still have to pick, and picking is the part that takes the time.

One post a week loses

The channels that reward volume won't wait for your free afternoon. The back catalogue keeps getting further behind.

Four steps, and two of them are yours.

01 — Import

Bring the recording in

Paste a YouTube link or upload the file. Sources up to four hours long, on the highest plan.

02 — Rank

Get a shortlist, not a document

Word-level transcription across the whole source, then every candidate moment is scored and ordered into up to fifteen clips.

03 — Frame

Check it the way a viewer sees it

Preview the clip in 9:16, 1:1 or 16:9. Choose the framing and the caption style. Trim if it cut a beat late.

04 — Export

One file per clip

MP4 up to 1080p, captions burned in, named so you know which clip is which.

The shortlist

It tells you which minute. And why.

Every suggestion is a real time range from your own transcript, scored on two independent tiers and sorted by strength. You get the ranking, a score label, a suggested title and a plain-English reason — so approving one takes seconds, and you can see exactly what was valued.

#1
0:42

The pricing mistake that cost me $40k

Opens with a concrete number, resolves inside a minute, and needs no setup from the rest of the episode.

Strong
#2
1:18

Why your first 100 customers don't come from ads

A contrarian claim stated plainly, then a specific alternative. The disagreement is the hook.

Strong
#3
0:36

One sentence that changed how I cold email

Self-contained story beat with a quotable last line, short enough to hold without a trim.

Solid

An example shortlist, showing the score label and reason that come with every suggestion. Nothing on it is invented: each clip is a time range from that recording's own transcript.

Framing

Shot wide. Posted vertical. Nobody's head gets cut off.

A sixteen-by-nine recording becomes a nine-by-sixteen clip that still looks intentional. Pick the fill that suits the footage — or, for a two-person conversation, keep both people in one vertical frame instead of cutting between them every four seconds.

  • Smart Fill
  • Fit Blur
  • Fit
  • Center Crop
  • Split-screen
  • Smart Fill Fills the vertical frame around the subject instead of leaving black bars.
  • Fit Blur Keeps the whole wide frame, with a soft fill behind it.
  • Center Crop, or fit the whole frame Two more options for footage where the subject isn't in the middle.
  • Three ratios, one project 9:16, 1:1 and 16:9 export from the same source and the same trim.
Captions

Captions land on the word, not near it.

Timing comes from the word-level transcript, so each line is placed where it was actually spoken. Choose a background treatment, tweak the style, and the preview you approved is the composition that gets exported. Karaoke style highlights each word as it's said.

  • Word-level timing Not a guess from a speech-to-text blob: each word carries its own start and end.
  • Editable before export If the transcript has a name wrong, fix it once and the captions follow.
  • Burned in The exported MP4 carries the caption, so it looks the same on every platform.

A transcript tool hands you a document. This hands you a decision.

What you're comparing Transcript tools SnipHawk
What comes back A transcript and a blade A ranked shortlist of up to 15 clips
Finding the moment You read, then guess Scored, ordered, and explained in a sentence
Framing Manual crop, clip by clip Smart Fill, Fit Blur, Fit, Center Crop, split-screen
Captions Auto-captions you re-time by hand Word-level timing from the transcript
Getting it out Download, then upload manually MP4 per clip, 9:16, 1:1 or 16:9, up to 1080p

What it won't do.

Worth saying plainly, because the opposite is usually what these tools promise.

It won't invent a clip

Every suggestion is a time range from your own recording. The scoring ranks and explains moments; it doesn't write them.

It won't post for you

SnipHawk exports the file. Where it goes, and when, is still your call.

It won't keep your footage

Working copies of your source are deleted on a schedule, and you can delete a project yourself at any time.

Your next recording is already the content.

Put the last one in. Take a month of posts out.

Request access