Diagnosis before treatment
Most 'make it better' tools re-render everything and hope. Hook Doctor names the specific problem first — a buried hook is a trim job, not a re-shoot.
Diagnose the first 3 seconds, then fix exactly that
Shorts live or die before the third second. Hook Doctor scores your opening 1-10, critiques the frames a viewer actually sees, names the ONE biggest problem — then applies the smallest fix that changes it: a tighter trim, punchier captions, a rewritten hook line. Before/after score included.
You type
This short is getting skipped — diagnose the hook and fix it.
…and Claude drives the whole pipeline below on your machine.
Why this beats a blank script box
It can actually see
Frame previews and vision critique mean the advice covers what the viewer's eye gets in second one — not just what the transcript says.
Measurable, not anecdotal
The 1-10 score before and after the fix is the deliverable. If the score didn't move, you iterate — or learn the problem is the format, and zoom out to a channel audit.
How Claude diagnoses and fixes a weak hook
-
Score the hook
tool_hook_analysis rates the opening 1-10 and returns a critique plus 2-3 rewritten hook lines — structured output, not vibes.
-
Look at what viewers see
preview_frames samples the opening moments so Claude can SEE the problem — a dead first frame, text arriving late, the subject too small — not just infer it from audio.
-
Name the one biggest problem
The diagnosis comes first, plainly: what's wrong and which single fix addresses it. Nothing renders until you agree.
-
Apply the smallest fix, re-score
Trim to the moment, burn viral-style captions, add zoom pulses, or re-voice the opening with a rewritten hook — then hook analysis runs again and you get the before/after score.
MCP tools it calls:
tool_hook_analysispreview_framescritique_videotool_trimtool_captions Connect it in 4 steps
-
Install ViralMint
Download the free, open-source desktop app (macOS / Windows / Linux) and open it. It bundles the whole pipeline — yt-dlp, Whisper, FFmpeg — and the MCP server.
-
Open the in-app MCP page
It shows your loopback endpoint and a bearer token, with a ready-to-paste Claude Code command and a JSON config snippet for Cursor / Claude Desktop.
-
Add the server to your agent
Paste the command. Your client loads ViralMint's tools, the built-in workflow recipes (this skill is one) and read-only resources like your balance and recent videos.
-
Ask for the outcome
Type the prompt above. Claude runs the recipe end-to-end against the local app and hands you the finished mp4 — no clicking through the UI.
ViralMint skills vs typical agent video tools
| What matters | Typical paid agent tools | ViralMint |
|---|---|---|
| Agent tools / skills | Behind a paid tier | Free & open-source (AGPL) |
| Where it runs | Their cloud | Your machine (loopback MCP) |
| Your footage & data | Uploaded to their servers | Stays local — only AI calls leave |
| Pricing | Monthly subscription | Prepaid, pay-per-render at ~cost |
| Output watermark | On free tiers | Never |
| Script quality | Prompt → generic script | Grounded in real breakouts + your data |
highlighted column = clearer fit.
FAQ
Frequently asked
What does the hook analysis actually check?
The spoken opening — is there a question, a stake, a promise in the first seconds? — scored 1-10 with a written critique and 2-3 rewritten hook lines. The frame preview adds the visual half: what's on screen while those words play.
Will it work on a video without narration?
Hook analysis needs a spoken opening, so pure-music videos get the visual critique path instead (frame previews + vision critique) — still useful for a dead first frame.
What fixes can it apply?
Trim straight to the moment, word-by-word highlight captions, zoom pulses on spoken words, a mild speed-up with pitch preserved, a re-voiced opening from a rewritten hook line, or a frame-anchored cover image. The smallest sufficient one goes first.
What if every one of my videos scores low?
That's a format problem, not a file problem — Hook Doctor hands off to the Channel Audit skill, which looks at your whole recent output and what overperforms for channels like yours.
Make {skill.name.toLowerCase()} from a chat
Install the free ViralMint desktop app, connect it to Claude Code or Cursor, and run this skill — no subscription, no watermark, your footage never leaves your machine.