i should never be sharing this but f*ck it
Gemini Omni + GPT Images 2 + tiktok is THE best combo for AI videos
i cracked the formula for generating videos that look hyper realistic & make your audience feel like "holy shit this person gets me"
i'm finally sharing my FULL system with you..
here's what you're getting:
- my prompting method for realistic voices (works every time)
- my realistic human movements & breathing trick (this is key for realistic videos)
- my exact method on how to make infinite length videos that maintain consistency
- how to get AI tools for dirt cheap (90% off)
RT + reply 'OMNI' and i'll send you the step-by-step system (must be following so i can dm)
i just finished the most valuable thing i've ever written.
"$0 → $10K/month with organic tiktok slideshows" — 17 pages. every mistake, shortcut, and system i wish someone handed me on day 1.
it'll be paid soon. right now it's free.
reply "PDF" + follow (so i can DM you)
I'm going to delete this soon
I’m 20 and my goal is to reach at least $2,000–3,000 per month in profit
I’ve narrowed it down to a few models
1. Selling digital products on Etsy
2. Amazon to eBay dropshipping
3. AI automation services
or is there anything better u would say?
i reverse engineered tiktoks duplicate detection system
heres exactly how they catch ur slideshows and how i bypass it to post the same images 500 times without getting flagged
tiktok doesnt manually review ur content
they run every image through perceptual hashing
its a math equation that turns ur pixels into a fingerprint
same images = same fingerprint = buried at 0 views
what doesnt work on its own anymore:
- cropping
- flipping
- mirroring
- reordering slides
- adding text overlays
- throwing filters on it
tiktoks 2026 ai sees through any of these as a single edit. they upgraded to deep learning models that read images the way u read them — by features not just pixels. one surface level edit isnt enough anymore. u need to stack multiple changes that actually break the fingerprint
heres how i do it at scale:
STEP 1 — ffmpeg batch processing
one command processes hundreds of images
- zoom 103%
- rotate 1-2 degrees random
- add 2-4px border random color
- slight brightness/contrast shift
stacking these together changes the hash way more than any single one
takes 30 seconds for 500 images
STEP 2 — subtle noise overlay
semi-transparent noise layer at 2-4% opacity
u literally cannot see it w ur eyes
but every single pixel value changes
breaks classical hash matching
STEP 3 — ai upscaling
for slideshows i want to spam hard i run them through topaz or real-esrgan
the ai doesnt just upscale, it generates new pixels through a gan
take a 1080p image, upscale to 4k, downscale back to 1080p
the pixels get completely reconstructed
original fingerprint is gone
this is the heavy artillery
STEP 4 — metadata spoofing
every file has metadata tiktok reads — creation date, device info, software signatures
unprocessed files look suspicious to the algo
spoof the metadata w exiftool or ffmpeg to look like the file came from a real iphone or capcut export
the algo trusts it instantly
my full workflow:
1. take winning slideshow images
1. batch process through ffmpeg (zoom, rotate, border, color shift)
1. add invisible noise overlay
1. run through ai upscaler for the ones i want to spam hardest
1. spoof metadata to look like a trusted source
one slideshow becomes unlimited unique variations
most ppl think the algorithm is random
its not. its a system
systems have rules
rules have exploits
do what u want w this information
i run one of these pages.
my friend runs the other 2.
total combined we've been able to generate +$700k in sales.
we're looking for health & wellness brands that are willing to be generous with really good commissions.
only DM if you're serious.