I just created a folder with 100 VIRAL YouTube formats you can copy and use for your own channel
Retweet & comment '1of10' below and I'll send it to your DM
(must follow to receive DM)
“every faceless youtube niche is saturated nowadays”
no mf
we’ve helped countless creators find untapped segments in "crowded" niches and print anyway
how?
by leveraging an AI system that finds outliers and reverse-engineers them to build winning scripts.
and today i documented the ENTIRE claude research workflow behind it...
here's what's inside:
→ the research objective most creators get wrong
(they research like journalists, hunting good information. but good info isn't what keeps someone watching. the angle, the structure, the data and the entry point are. this resets what you're even looking for)
→ the 3 research mistakes quietly killing your retention
(researching broadly instead of strategically, ignoring the outliers already beating their channel 5-10x, and researching the topic instead of the audience, the one that makes viewers leave at second 20)
→ phase 1: outlier transcript analysis (8 min)
(paste a viral outlier's transcript, a 5-point prompt extracts the structural map of why it won: hook structure and the trigger it fires, where the FIRST open loop sits, the pacing pattern, every mini payoff and any payoff void, the grand payoff, and the vocabulary fingerprint)
→ phase 2: competitor channel analysis (4 min)
(feed 3-5 channels' last 15-20 uploads and view counts, claude maps what's saturated, what's overperforming, and the single-channel outlier topic that proves demand with the lane still open)
→ phase 3: search intent mapping (4 min)
(the gap between what a viewer types and what they actually want, turned into 3 ready hook angles, contradiction, dramatic stakes, problem agitation, before you write a word)
→ phase 4: script brief assembly (4 min)
(all 3 phases folded into one engineering document: hook, entry point, 5-6 mini payoffs, transitions, grand payoff, viewing experience. the thing that produces a tight script on ANY model)
→ calibrating the analysis per format
(the questions you ask a true crime transcript aren't the questions you ask a celebrity doc, so the same 5-point breakdown tunes to whatever you actually write)
all backed by 7,000+ scripts across 50+ niches and 150M+ views generated for clients with scriptwriting.
it's free for the next 24h, then it goes back into our internal vault.
like + comment "SCRIPTS" and i'll send it over
(must be following + RT for priority access)
just got an average retention rate of 64.2% in a niche where it's 14.8%.
→ both from the same niche
→ same upload strategy
→ one variable: the hook
at normal speaking pace, 50 words is 20 seconds. youtube's own data shows 1 in 5 viewers leaves in that window. and most faceless creators are burning those 50 words saying nothing.
so i broke down the ENTIRE hook engineering system behind that 64.2% number, the same system i forced into claude opus 4.8...
here's what's inside:
→ the 2 hook frameworks
(character-concept-stakes vs target-transformation-stakes, and why running the wrong one for your content type is the exact thing that collapses retention at 20 seconds)
→ the 4 hook types by format
(in media res, the contradiction hook, the direct question, problem agitation, and the precise moment each one fires)
→ the psychological triggers
(curiosity gap, forbidden knowledge, dramatic stakes, specificity bias, loss aversion, the levers that make a viewer unable to scroll past second three)
→ the 6 hook killers
(credential dumping, jargon overload, generic openings, weak stakes, the silent retention leaks bleeding out of EVERY script you've written)
→ the templates
(real bad-to-good rewrites for every framework and every hook type, so you can see the fix, not just the rule)
→ the claude opus 4.8 skill file setup
(how all of it loads into claude via FacelessOS, 16 skill files that write the hook for you instead of you rewriting 80% of the output)
all backed by 7,000+ scripts across 50+ niches, 150M+ views, and a 64.2% vs 14.8% live retention comparison from the same niche.
like + comment "SCRIPTS" and i'll send it over
(must be following + RT for priority access)
will stop sending these out in 24h
@DavidNortes_ This has been very inspiring for me. And because of that, I intend to document my starting jourey on X as well.
Congratulations to you two
i've spent 100+ hours breaking apart what youtube's gemini algorithm actually cares about
and today i've decided to document ALL my findings...
every signal taken from your script, analytics and what viewers do AFTER the video, plus the exact scripts that force the algo into loving your faceless channel…
here's EXACTLY what's included inside that guide:
→ what actually changed
(youtube stopped sorting by broad topic, gemini now builds a granular interest profile per viewer and matches with surgical precision, which means you only compete against channels with your EXACT content signature now, and a generic signature gets you buried no matter how good the edit is)
→ how gemini reads you across 3 layers
(transcript analysis for expertise level, editorial POV and whether it scores as default-AI output, visual analysis for footage-to-script match and production value, and viewer-behavior signals for what they do AFTER the video ends)
→ the 4 signals it actually weighs
(CTR paired with early retention so clickbait that delivers is fine but clickbait that doesn't is now actively PENALIZED, retention curves at a flat 50% beats a spike that crashes by minute 5, satisfaction signals from post-view behavior, and session contribution measuring whether your video makes them watch MORE youtube)
→ the 4 script builds that move those signals
(the CTR-retention bridge that reiterates the hook's promise in the first 2 minutes, the retention architecture running the first 50 formula in minute 0-1 on a 1-5 context-action ratio, the satisfaction build that over-delivers early and answers objections in-script, and the session moves that hand viewers their next video)
→ everything off the video gemini ALSO reads
(150-300 word descriptions that act as its second document, 10-15 tags, intent-signalling chapters, plus the channel-level scores: niche clusters, upload consistency, returning-subscriber % and content depth getting more sophisticated over time)
→ (BONUS) the 5-minute gemini audit
(a pre-upload checklist you run before you publish, plus the 3 inauthentic-content patterns that terminated 16 channels in january 2026 and exactly how to dodge structural uniformity, AI language-pattern matching and a low originality score)
all backed by my experience of generating 150M+ views for clients with youtube scriptwriting
like + comment "GEMINI" and i'll send EVERYTHING your over
(must be following + RT for priority access)
@fyreinteractive@fyreinteractive please it seems the attached images are that of the prior article and not for "the psychology behind faceless youtube channels that print $10k-$100k/month"
Trying to find out if it's on purpose or just a mistake?
Thank you
IT’S OVER…Photoshop and Canva are cooked.
We just made the 1of10 thumbnail generator 100x more powerful…
And it’s INSANE.
You can now generate AND fully edit your thumbnails with AI, all in one place.
- Magic Edit: scribble on any part of the thumbnail, type what you want changed, and DONE.
- Magic Eraser: highlight anything you don't want and it's GONE, background filled like it was never there
- Face Swap: drop your own face onto anyone in the thumbnail in one click
- 4K Upscale: We fine tuned our upscale feature to turn any thumbnail crisp and sharp, built for YouTube's new 4K standard.
Plus a ton of improvements across the generator that make every thumbnail come out even better than before.
I've been playing with this all week and honestly can't believe what it does. This is a Photoshop killer. So hyped for you all to try it.
To celebrate, we're giving one person a FREE year of 1of10. Retweet and comment “1of10” to enter, we'll pick a winner in 48 hours.
i've been using claude + nexlev MCP to reverse-engineer the faceless channels blowing up as we speak...
then writing scripts with opus 4.8 based on the advanced data i pull
same system i used to generate 150M+ YT views
so i've decided to document the A-Z blueprint:
> finding niches actually worth your time
nexlev surfaces sub-50k channels suddenly pulling 200k+ views. you want $7+ RPM and under 5 weak channels already in there. the same video pays 10x in a $30 niche vs a $3 one, skip the cheap ones.
> ripping the outliers
feed the winner's transcript into claude, pull the hook, the pacing, where the curiosity gaps open and close. you're not copying the topic. you're copying the structure that held people to the end.
> writing so gemini serves you right
nobody talks about this. gemini reads your script to decide who sees the video. "real estate is a good investment" pulls broke browsers. "cap rate compression in tier-1 markets" pulls people with portfolios. same topic, 10x the RPM. the words pick the audience.
> staying off the slop pile
the 8 things gemini's quality filter scores you down for. summarizing intros, flat sentence rhythm, hedging, context dumps. kill them before you upload.
> publish + let it self-correct
push it with the SEO stuff, then 48 hours later pull your CTR, retention and RPM back in and fix the next script off real numbers.
all backed by 7,000+ scripts responsible for $5M+ across 50+ niches
and for 24h, it can be EVERYTHING yours (for free)
like + comment "SCRIPTS" and i'll send it
(must be following + RT for priority access)