Pinned Tweet
- new accounts can’t go viral anymore
- Instagram doesn’t work like tiktok
- etc etc etc
yes it can good sir, yes it can.
golden rule to remember: virality can’t be invented, it can be engineered
i.e. as long as you pick something that’s already gone viral before, chance of higher.
and it’s made easy for you with @handlerai_, find all winning viral outliers there.
‼️ only for apps and SaaS, ecom not covered 🫡
- Instagram doesn’t work like tiktok
- etc etc etc
yes it can good sir, yes it can.
golden rule to remember: virality can’t be invented, it can be engineered
i.e. as long as you pick something that’s already gone viral before, chance of higher.
and it’s made easy for you with @handlerai_, find all winning viral outliers there.
‼️ only for apps and SaaS, ecom not covered 🫡
15
8
117
7,952
This week on TikTok - Outliers spotted by Handler to market your product 🔥
✅ categories: entertainment, lifestyle, b2b and other apps
> 📺 this week in entertainment: a retro music player post pulled 789.9k views from an account that usually gets 3.5k. six breakouts, all scosame play.
> 💃 this week in lifestyle: a christian devotional bot post pulled 2.9m views from an account that usually gets 356. six breakouts on the board, all scored so you can dupe what worked.
>👨💼 b2b tiktok this week: an ai side hustle clip stacked 1.6m views from an account that usually gets 5.5k. six breakout posts, all scored so you can run the same play.
> 🌈 utility apps had a wild week: a bluetooth device finder post pulled 6.7m views from an account that usually gets 1.3k. six fresh breakouts on the board, all scored so you can dupe the play.
Swipe the slides to see them category by category ➡️
full breakdown here: gethandler.ai/outliers/this-…
✅ categories: entertainment, lifestyle, b2b and other apps
> 📺 this week in entertainment: a retro music player post pulled 789.9k views from an account that usually gets 3.5k. six breakouts, all scosame play.
> 💃 this week in lifestyle: a christian devotional bot post pulled 2.9m views from an account that usually gets 356. six breakouts on the board, all scored so you can dupe what worked.
>👨💼 b2b tiktok this week: an ai side hustle clip stacked 1.6m views from an account that usually gets 5.5k. six breakout posts, all scored so you can run the same play.
> 🌈 utility apps had a wild week: a bluetooth device finder post pulled 6.7m views from an account that usually gets 1.3k. six fresh breakouts on the board, all scored so you can dupe the play.
Swipe the slides to see them category by category ➡️
full breakdown here: gethandler.ai/outliers/this-…
0
0
0
87
An Nayal retweeted
can a travel app grow on TikTok without paid?
one travel app we work with -> 22M organic views
travel is the easiest niche to get views in and one of the hardest to turn into installs. everyone saves the video. few download the app.
so we brief for the download, not the save 🫡
ganic UGC for consumer apps, from creator matching to the weekly call on what to kill and what to scale.
agniverse.co/media
one travel app we work with -> 22M organic views
travel is the easiest niche to get views in and one of the hardest to turn into installs. everyone saves the video. few download the app.
so we brief for the download, not the save 🫡
ganic UGC for consumer apps, from creator matching to the weekly call on what to kill and what to scale.
agniverse.co/media
4
1
4
606
calendly charges per seat for what is basically one google api call
free/busy in -> open slots out -> create the event with a meet link
so i built it myself and put my whole team on it
next.js + supabase + one google service account, deploys on vercel in under 20 min
open sourcehere 👇
github.com/Git-Agni/open-…
free/busy in -> open slots out -> create the event with a meet link
so i built it myself and put my whole team on it
next.js + supabase + one google service account, deploys on vercel in under 20 min
open sourcehere 👇
github.com/Git-Agni/open-…
done, it's open source now ✅
iOS phone farm - register real iPhones, watch + control them live in the browser, schedule tiktok on a postgres-backed scheduler.
free, self-hosted, apache-2.0.
- git: github.com/Git-Agni/prod-…
- DIY steps: gethandler.ai/ios-farm
1
0
0
187
per event: its own hours, buffers, notice, max per day, custom questions, add guests on or off
no seats, no plan, no upsell. one admin for the whole team
no seats, no plan, no upsell. one admin for the whole team
1
0
0
22
we built it at Agniverse to replace calendly for our team. democratizing software for all 🫡
0
0
0
13
An Nayal retweeted
- new accounts can’t go viral anymore
- Instagram doesn’t work like tiktok
- etc etc etc
yes it can good sir, yes it can.
golden rule to remember: virality can’t be invented, it can be engineered
i.e. as long as you pick something that’s already gone viral before, chance of higher.
and it’s made easy for you with @handlerai_, find all winning viral outliers there.
‼️ only for apps and SaaS, ecom not covered 🫡
- Instagram doesn’t work like tiktok
- etc etc etc
yes it can good sir, yes it can.
golden rule to remember: virality can’t be invented, it can be engineered
i.e. as long as you pick something that’s already gone viral before, chance of higher.
and it’s made easy for you with @handlerai_, find all winning viral outliers there.
‼️ only for apps and SaaS, ecom not covered 🫡
15
8
117
7,952
An Nayal retweeted
we had to pause a client's UGC because it worked too well
2.5M views in 4 days
one video hit 1M from a brand new account. 3 posts, 87 followers
the app is still waiting on App Store approval, so nobody could download it 😂
the plan was to start before launch, test angles and d on launch day
we found what works a little early
relaunching the second Apple says yes 🫡
2.5M views in 4 days
one video hit 1M from a brand new account. 3 posts, 87 followers
the app is still waiting on App Store approval, so nobody could download it 😂
the plan was to start before launch, test angles and d on launch day
we found what works a little early
relaunching the second Apple says yes 🫡
3
2
30
2,437
#tb to this banger 😂
doesn’t matter the location, if I hear kala chashma the desi comes out
0
0
0
150
An Nayal retweeted
This week on TikTok: one entertainment app post exploded to 231x its normal baseline 🔥
📻 A retro music-player post pulled 789.9K views from an account that usually gets around 3.5K.
That’s exactly the kind of signal Handler is built to catch.
Instead of scrolling TikTok and k, Handler surfaces the posts already massively outperforming in your category.
Then it breaks down the hook, angle, format, and creative pattern so you can turn proven content into ideas for your own app.
You’re not copying the video.
You’re stealing the reason it worked.
This week, Handler found 6 entertainment breakouts worth studying.
→ Find the outliers
→ Understand what made them win
→ Adapt the hook to your product
→ Ship faster
→ Test from a proven starting point
Less guessing. Better creative inputs.
→ gethandler.ai/outliers/this-…
📻 A retro music-player post pulled 789.9K views from an account that usually gets around 3.5K.
That’s exactly the kind of signal Handler is built to catch.
Instead of scrolling TikTok and k, Handler surfaces the posts already massively outperforming in your category.
Then it breaks down the hook, angle, format, and creative pattern so you can turn proven content into ideas for your own app.
You’re not copying the video.
You’re stealing the reason it worked.
This week, Handler found 6 entertainment breakouts worth studying.
→ Find the outliers
→ Understand what made them win
→ Adapt the hook to your product
→ Ship faster
→ Test from a proven starting point
Less guessing. Better creative inputs.
→ gethandler.ai/outliers/this-…
0
1
3
130
An Nayal retweeted
tiktok used to be 80-90% of the views and conversions for app UGC.
not anymore 😳
@mattgittleson (founder HardLaunch, 2.5B views in the last 6 months and overall legend) on 🥫The App Sauce Pod:
"instagram is now the priority channel for most apps. facebook is quietly working en place over the last 6 months"
we're seeing the same on our end:
- a prosumer B2B app cracked on facebook. not tiktok, not IG, not youtube
- fresh IG accounts went viral in 3 days for a client whose app wasn't even live yet. had to pause the campaign
tiktok isn't dead. slideshow niches still print there.
but if your UGC plan is just "tiktok marketing", you're skipping the bigger channel
pod with Matt the legend coming out soon 🔥🫡
not anymore 😳
@mattgittleson (founder HardLaunch, 2.5B views in the last 6 months and overall legend) on 🥫The App Sauce Pod:
"instagram is now the priority channel for most apps. facebook is quietly working en place over the last 6 months"
we're seeing the same on our end:
- a prosumer B2B app cracked on facebook. not tiktok, not IG, not youtube
- fresh IG accounts went viral in 3 days for a client whose app wasn't even live yet. had to pause the campaign
tiktok isn't dead. slideshow niches still print there.
but if your UGC plan is just "tiktok marketing", you're skipping the bigger channel
pod with Matt the legend coming out soon 🔥🫡
6
2
33
2,562
instinct will IPO before anthropic at this rate
Instinct is currently raising @ $30b.
$100M → $500M → $2.5B → $10B → ~$30B in sub 1y.
If you're still not inside the private market this should be a good signal to pivot aggressively.
Wealth creation is going to be heavily skewed toward private investors over the coming decade.
0
0
0
315
An Nayal retweeted
after pushing out more content, i can confidently vouch for the accuracy of the analysis below.
tiktok and motion control doesn't work.
tiktok and motion control doesn't work.
I may have figured out how TikTok detects duplicate videos, and why AI-generated videos get flagged while slideshows often don’t.
My theory: TikTok isn’t just matching pixels. It’s matching MOTION.
Here’s what we’ve observed from our experiments with TikTok farming and A thread 🧵
1/ The problem with AI motion control
We’ve been experimenting with Kling AI’s Motion Control.
For context, Motion Control lets you take an existing video, extract the movements, and recreate those exact movements using an entirely different AI-generated character.
Different face. Different body. Different visuals.
But almost identical movements.
And here’s where things get interesting.
2/ The first experiment
We took existing videos that had already been posted on TikTok.
Used Kling Motion Control to recreate them with entirely different AI characters.
Then uploaded those newly generated videos.
Result? They repeatedly got flagged as unoriginal or restricted.
Even though the actual person in the video was completely different.
3/ We thought it might be AI detection
So we experimented with removing AI-related metadata and identifiers.
Same outcome.
That made me question whether TikTok was actually detecting the AI-generated character at all.
What if it was recognizing something that stayed constant between the original video and its AI recreation?
The MOTION.
4/ Here’s my actual hypothesis
TikTok may be using temporal video fingerprinting to recognize duplicate content.
Instead of only comparing what individual frames look like, it may be comparing how objects move between frames.
Think:
• Body movements
• Hand gestures
• Head movements
• Facial expressions
• Timing and sequence of actions
Essentially, a motion signature for each video.
Two videos could have completely different characters but still share an extremely similar motion signature.
5/ Why Kling Motion Control makes this interesting
Kling can recreate the movements of an existing video with remarkable accuracy.
That’s literally what Motion Control is designed to do.
But if TikTok fingerprints the underlying movement patterns, changing the character might not be enough to make the video meaningfully different.
The pixels change.
The motion remains.
And that could explain why these AI recreations repeatedly get flagged.
6/ The second experiment is what really got me thinking
We tried something different.
Instead of taking the motion reference from an existing TikTok video, we recorded our OWN original footage.
Then used that footage as the reference for Kling Motion Control.
Same AI generation workflow.
Same underlying technology.
But the results were different.
These videos were significantly less likely to get flagged in our testing.
The key difference?
The motion itself was original.
7/ Now here’s where SLIDESHOWS come in
We’ve also noticed something interesting across our TikTok experiments.
Replicated slideshow formats often perform much better in terms of avoiding unoriginal-content restrictions than replicated videos.
Even when they’re heavily inspired by existing slideshows.
Why?
My theory is that slideshows have a fundamentally different fingerprinting problem.
8/ Videos contain a temporal dimension that slideshows largely don’t
A video has continuous movement.
Potentially 30 or 60 frames every second.
Those frames capture a sequence of movements over time.
A slideshow, meanwhile, consists primarily of static images.
TikTok can still analyze image similarity, text, audio, and other signals.
But there isn’t the same continuous motion signature to compare.
That could make motion-based duplicate detection much more effective for videos than for slideshows.
9/ My theory on the underlying technology
🧵👇🏽
3
1
4
909
An Nayal retweeted
we put in scratch gen to resolve the issue, allowing you to generate respective clip from scratch by using your own motion! 🫡
@consumerxai so jean phil video style does not work ? and what did you do with dupes on your saas to solve that ?
0
1
2
345
An Nayal retweeted
how to tell if a tiktok video is sparked and has views boosted as an ad?
- right-click anywhere on the page → view page source
- or press Ctrl + U (Windows/Linux) / Cmd + Option + U (Mac)
- this opens the original HTML document in a new tab.
- use Ctrl + F to search for "isAd".
ed.
voila 🫡
- right-click anywhere on the page → view page source
- or press Ctrl + U (Windows/Linux) / Cmd + Option + U (Mac)
- this opens the original HTML document in a new tab.
- use Ctrl + F to search for "isAd".
ed.
voila 🫡
0
1
11
921
An Nayal retweeted
we stopped handing out business cards at conferences
now people scan a QR, type their name, and fly a rocket to mars
they land on a boarding pass with my links + contact card. i land in their inbox 📥
done, open sourced this one too ✅
rocket, horse or a solid-gold gorilla - el
fork it, change one config file, print your QR - free here👇
github.com/Git-Agni/fligh…
now people scan a QR, type their name, and fly a rocket to mars
they land on a boarding pass with my links + contact card. i land in their inbox 📥
done, open sourced this one too ✅
rocket, horse or a solid-gold gorilla - el
fork it, change one config file, print your QR - free here👇
github.com/Git-Agni/fligh…
done, it's open source now ✅
iOS phone farm - register real iPhones, watch + control them live in the browser, schedule tiktok on a postgres-backed scheduler.
free, self-hosted, apache-2.0.
- git: github.com/Git-Agni/prod-…
- DIY steps: gethandler.ai/ios-farm
1
1
1
430