Say the topic. Make them curious. Flip it.
Say what it's about in line one, in voice, picture and text. Then a stat or pain. Then the twist.
I studied the 6 reels you sent. Here is the playbook, what it changes in our 2 samples, and 4 things I need from you.
Everything else on this page is detail under these three.
Say what it's about in line one, in voice, picture and text. Then a stat or pain. Then the twist.
Every caption must pass 3 tests: you can picture it, it can be proved true or false, and no one else could say it.
A cut, a zoom, a new part of the diagram, a new number. And a voice: all 6 reels talk, none are text-only.
Pick a sample. Switch between the old hook and the new one. Turn each layer on and off to see what it adds.
Tap a rule. Each one says where it came from and what changes in our videos.
Each block is one shot. Striped blocks are longer than the limit. Drag the slider to set the limit.
Our diagrams already move inside each beat. This counts only the big changes (new shot, new scene, zoom). The 6 reels make one every 3 to 6 s; the fastest every 3 s.
The new headphones-and-laptop durian becomes the face of every video. It does things. Tap an action.
I'll redraw this art as a rig (moving eyes, arms, laptop, headphones that glow on the beat) so it can act in 2D and 3D. Same character as today, now with headphones and a laptop.
Made before this playbook. Accurate and clean, but they break rules 1, 4, 5 and 11. Next versions fix that.
Answer on the agenda card. My pick is marked on each one.
All 6 reels are voice-led. Rule 1 needs the topic in the audio too.
A name makes it a character people remember and comment to.
"Comment a word and I'll send you X" gets the most comments, but someone has to send X. Free tools don't do it automatically.
If yes, I write it into the repo's CLAUDE.md, so every future video follows it. Then I rebuild both samples with it.
Then I rebuild both samples with the new hook, voice and acting durian.