Phase 1: The "Un-Detectable" Script

The biggest mistake new creators make is pasting a raw ChatGPT script into YouTube. The algorithm detects the robotic cadence and kills the reach.

Step 1: The Prompt. Do not ask for a script. Ask for an outline. Use this prompt:

"Act as a documentary storyteller. Write a 300-word script about [Topic] that focuses on a mystery or conflict. Use short sentences. Do not use words like 'unleash', 'dive in', or 'tapestry'."

Step 2: The Scrub. Take that text and run it through a Humanizer. This varies the sentence structure (Burstiness) so it sounds like a human wrote it.

Tool: Undetectable.ai

Use the "University" readability setting for best results.

Get Tool →

Phase 2: The "Breathing" Voiceover

Old Text-to-Speech (TTS) sounds monotonic. Modern "Generative Voice" adds breaths, pauses, and pitch changes based on context.

Configuration Settings: When using ElevenLabs, do not leave settings at default. Use these values for a documentary style:

Stability: 45% (Lower stability = more emotion/variation)
Similarity Enhancement: 75% (Keeps the voice consistent)
Style Exaggeration: 10% (Keeps it grounded)

Tool: ElevenLabs

Use model "Eleven Multilingual v2" for best flow.

Get Tool →

Phase 3: The "Generative" Edit

You have the audio. Now you need visuals. You have two paths:

Path A: Long Form

Use InVideo AI. Upload your script. It will scrape stock footage (iStock/Shutterstock) and build the timeline for you automatically.

Path B: Short Form

If you have a long video already, use OpusClip. It will slice it into 60-second viral clips with "Alex Hormozi" style captions.

Video Tools

Compare OpusClip vs InVideo features.

Compare →