Content Creator

Background

The Challenge

Too much inspiration, too little structure: 100+ Notion cards, 3-5 reference videos, 2-3 long reads — there was no way to judge by gut which points belonged in the first 30 seconds and which belonged at the end. First drafts averaged 4-5 rewrites; prep for a single video took 2 days.

The Approach

He fed all sources — Markdown-exported Notion cards, YouTube URLs, article URLs — into MindLM in one go, letting the AI cluster the scattered material into a thematic map. Then in the editor he reorders themes, adds hook lines, and annotates each branch with “what reaction the viewer will have after this section.” The map exports directly as a Markdown script skeleton.

Workflow Walkthrough

1. 1. Bulk-import all idea cards

   Export 100+ Notion cards as a single Markdown file, drop it into MindLM. The AI auto-clusters into 6-10 themes — replacing the old manual sorting of which cards belong together.

2. 2. Merge in video transcripts and articles

   Paste 3-5 YouTube URLs and 2-3 article URLs. MindLM generates a map for each in 60-90 seconds. He hangs those reference maps as sub-branches under the master map — every reference point now stands shoulder-to-shoulder with his own cards.

3. 3. Add a 'viewer reaction' note per theme

   On every top-level theme he annotates the expected viewer reaction — laugh / surprise / save / pause-to-note. If two adjacent themes predict the same reaction, one must be broken out — a pacing rule he learned the hard way from 4-5 prior rewrites.

4. 4. Export as a script skeleton

   Export the master map as Markdown. The first two H2s are hooks, the middle 5-7 H2s carry the body pacing, and the last H2 closes. Each H3 keeps a one-sentence content directive — full word-for-word script is filled in only 1 hour before recording, to avoid locking the language too early.

Results

• Prep time per video cut from 2 days to 1
• Pacing-rule ordering lifted first-30s retention by 18-22%
• Monthly output rose from 4 to 5-6 videos with less prep per video

By the Numbers

Once it was running

Under the new flow, prep for one video shrunk from 2 days to 1. The bigger surprise was viewer data: in the second month he noticed that videos ordered by the pacing rule had 18-22% higher retention in the first 30 seconds, with average watch time up by roughly a minute. Three months in, he raised monthly output from 4 to 5-6 videos while still spending less prep time per video.

"MindLM turns my ideas into structure, then turns structure into pacing. I used to think creation was talent. Now I believe it is process."

Key Takeaways

• 1.Don't pipe inspiration into a script directly — cluster into themes first, then decide order. It's the only way to land both information density and pacing.
• 2.Annotate each theme with the expected viewer reaction: adjacent themes must not predict the same reaction, or the pacing will collapse.
• 3.Skeleton before script: lock H2/H3 first, then write the actual lines only 1 hour before recording — preserves the live feel.

Try this tool yourself