456
Prompt

YouTube Script Prompt (Retention-Optimized)

Generate a hook-driven script tied to retention beats.

1 min read

/ quick answer

Use as the core step of the YouTube script automation pipeline once a hook is chosen. Generate a hook-driven script tied to retention beats.

Generate a hook-driven script tied to retention beats. Use as the core step of the YouTube script automation pipeline once a hook is chosen. Copy the prompt below, swap the bracketed variables for your own context, and run it in any capable model. This prompt node is part of the Onexial knowledge graph and links to related concepts, workflows and tools below.
Context
Use as the core step of the YouTube script automation pipeline once a hook is chosen.
Prompt
You are a top YouTube scriptwriter. Write a {{length_minutes}}-minute script on:

Topic: {{topic}}
Chosen hook: {{hook}}
Audience: {{audience}}

Structure:
1. Hook (0–15s): cold-open, no logo, no intro.
2. Promise (15–30s): what they'll get + why they should stay.
3. Body: {{beat_count}} beats, each with a mini-cliffhanger to the next.
4. CTA: tied to the topic, not generic "like and subscribe".

For each beat, include: [B-ROLL: ...] cues.
Example Output
[0:00] "Most AI agents fail before they make a single API call."
[0:12] In the next 6 minutes I'll show you the 3 reasons — and the fix used by every working agent stack.
[B-ROLL: terminal with failed agent loop]
...
Related Workflow
Related Tool Stacks
/ frequently asked

What does the YouTube Script Prompt (Retention-Optimized) prompt do?

Use as the core step of the YouTube script automation pipeline once a hook is chosen.

Which AI models work with this prompt?

It is model-agnostic: it works with any capable general model. Replace the bracketed variables with your own context before running it.

What output should I expect?

[0:00] "Most AI agents fail before they make a single API call." [0:12] In the next 6 minutes I'll show you the 3 reasons — and the fix used by every working agent stack. [B-ROLL: terminal with failed agent loop] ...