how I trained Claude to write $50,000 worth of YouTube scripts in 2 minutes
the real breakdown (with prompt stacks, system setup, and quality control🧵
first, context
i run faceless channels
multiple
different niches
scripts are the bottleneck
and hiring more writers doesn’t scale, it just multiplies inconsistency
so i built a system to train AI to write scripts that sound human, structured etc
it took weeks, now it prints
let’s be clear
i treat it like a junior writer on payroll
that means:
i give it a role
i give it a framework
i give it examples to mimic tone + pacing
i set quality metrics
the setup
my pipeline looks like this:
idea input → format prompt → tone injection → outline pass → script build → human QC
each part has its own micro-prompt
not one giant paragraph
the magic isthe stack
prompt stack example (simplified)
Step 1: role assignment
“you are a professional youtube scriptwriter who writes like [channel name],
focusing on emotional pacing, retention hooks, and short-sentence rhythm.”
Step 2: outline creation
“create a 7-section outline with cliffhangers every 60 seconds. each section should end with tension.”
Step 3: script expansion
“expand section 1 into a script, using sensory language, timestamp pacing, and conversational narration.”
Step 4: tone reinforcement
“rewrite in the tone of a human creator, mix confidence, sarcasm, and curiosity.
remove robotic phrasing. break rules when needed.”
Claude remembers tone consistency
it keeps human pacing across long scripts (8-10 mins)
you can literally feed it 3 sample scripts from your best writer, and it starts mimicking flow + structure instantly
i added a human QC layer
my editor doesn’t rewrite, he just checks
5-minute review = 10x better script
before:
2 writers = 4 scripts/week
after:
Claude = 12 scripts/day, with QC done by one human
that’s 50k worth of deliverables per month if done manually
cost = $20/month Claude subscription + QC time
you can join Faceless Success to get my actual prompts i use for AI model 🤷♂️
ROADMAP TO $22,000/MONTH USING MULTIPLE FACILESS CHANNELS 🧵
PHASE 1: THE FOUNDATION ($0 → $2K/month)
Goal: Build one proof-of-concept channel that actually works
Key steps:
Pick a content model that already prints views
Study top channels. Steal the format, not the topic
Do 30+ uploads to learn what retention, CTR, and satisfaction actually mean in your niche
Automate using low-cost tools
Get monetized.
You’re learning systems logic here: inputs, outputs, optimization
Once you can generate $5–$10 RPM with consistent retention and clicks, you’ve built your first machine.
PHASE 2: SYSTEMIZATION ($2K → $8K/month)
You’ve proven your workflow works
Time to clone it
What changes here:
You outsource scriptwriting and editing
You document your entire workflow
You create a feedback loop
You test new channels with the same format, same niche
how to build your first faceless YouTube team (without losing money) 🧵👇
you think hiring a team will magically fix your workload but it won’t
if you don’t already have a proven content structure, you’re just paying people to experiment
you only hire once you’ve done at least 10 videos yourself and know:
what works
what doesn’t
what can be outsourced
your team is there to scale systems, not figure them out for you
In Faceless Success, you can have access to my team but that is not the only thing you get
- AI YT model
- INSANE affiliate program
- group coaching calls ++