7.0M views
“Crazy texts with my MIL”
MusicTikTokTracked July 18, 2026
How an AI song app got 27M views on mother-in-law drama
Mozart AI is an AI song generator: you type text or drop a photo, pick a genre, and it makes a full song with lyrics. It ships one-tap sharing to TikTok, Reels and Shorts, its tracks are royalty-free, and its store listing claims 10M+ creators.
7.0M views
4.1M views
1.7M views
The campaign's three biggest hits. We take them apart below.
The thesis
Mozart AI stopped selling "make songs with AI" and started selling mother-in-law drama. The song became the audio, and the views went up 100x.
The short answer
In one month (Jun 15 to Jul 14, 2026) Mozart AI's creator campaign pulled 27.4M views across 1,298 TikTok posts, peaking at 6.97M views in a single day. It ran on 16 creator accounts posting about 43 times a day. The winning move was not a product demo: creators fed dramatic text conversations, a crazy mother-in-law, a traumatizing dating app, into Mozart, got a song back, and posted them as serialized sagas over mundane kitchen b-roll. The posts that flopped were the ones that actually advertised the app.
27.4Mviews in one monthpeaked 7.0M on Jul 8, 2026
1,298posts16 creators, ~43 a day
21Kaverage views per postbut the median was only 275
3.56%engagement rate942K likes total
How it grew
A volume game with a fat tail
It did not grow because every post worked. It grew because a small team posted enough drama-hook variations that a few caught fire.
This was a volume game with a fat tail. 16 creator accounts posted 1,298 times in 30 days, roughly 43 posts a day. The median post got just 275 views. But six posts crossed a million, and those six carried most of the 27.4M. Growth did not come from every post working. It came from posting enough drama-hook variations that a few caught fire, then serializing the winners.
16creator accounts
43posts a day
30active days
The hit rate
Of 1,298 posts, only 6 crossed a million views and 21 crossed 100K. More than 1,078 got under 1,000. The median post got 275 views. A handful of drama hits carried the whole 27.4M.
What wins, what flops
The product was the worst hook
These are real posts from the same campaign, same app behind them, captured after days online (not fresh posts). The difference is the hook. The winners open a mother-in-law fight you have to finish. The flops lead with the song. The takeaway is blunt: the product is the worst possible hook.
4.1M views
“Part 2 (prequel): Crazy texts from my MIL”
1.7M views
“Crazy texts with my MIL before I give birth”
1.6M views
“Part 2: texts with my crazy MIL about giving birth”
767K views
“reasons hinge is traumatizing me pt. 1”
651K views
“Part 4: Text between my husband & his mom”
And what died: the posts that sold the app
Same app, same creators, near-zero views. Watch one: the hook is about the product, not a story.
1 view
“could this be the song of the summer???”
This one screen-records the Mozart app making a track and asks if it is a hit. The product is the entire video. It got 1 view. The 7M-view post never shows the app at all.
1 view
“turning my sister's embarrassing texts into a gospel song”
It has drama, a sister's texts, but the hook leads with the song, so the promise is the product, not the fight. 1 view.
2 views
“your brother makes music? that must be so cool”
Leads with the music angle instead of a conflict. The winners open with a fight and mention music last, if ever. 2 views.
The format
What they actually post
There are three formats, and the split tells the whole strategy. 61% is a text conversation over ordinary footage, the workhorse. 29% is a creator narrating to camera. Only 10% is the one format where the product actually appears. Here is what each looks like, with a real example from the campaign.
Wall of Text 61%Talking Head 29%Hook + Demo 10%
61%
Wall of Text
7.0M viewsA screenshot text conversation over mundane b-roll (a kitchen, hands, a car). The drama unfolds message by message while nothing happens on screen. This is the workhorse: 6 of every 10 posts, and every top winner.
e.g. “Crazy texts with my MIL”
29%
Talking Head
209K viewsA creator talks straight to camera about the drama. Higher trust because there is a real face, but slower to make and easier to clock as content. Used when the story needs a narrator.
e.g. “the reality of having a DJ sister”
10%
Hook + Demo
68K viewsThe only format where the product shows up. A creator replies to a comment asking how they made the song, then demos Mozart. The app appears here, after the drama already earned the attention, never before it.
e.g. “how do you make this song?”
The engine
One hook, then a saga
The real repeatable move: take a proven opener and serialize it into parts. Each family below is one idea, multiplied.
“crazy texts with my mil”
best: “Crazy texts with my MIL”
3parts
137Ktop likes
“reasons hinge is traumatizing me”
best: “reasons hinge is traumatizing me pt. 1”
10parts
21Ktop likes
“the reality of having a...”
best: “the reality of having a DJ sister”
12parts
116top likes
The cadence
When and how often they post
Posts per day~43/day avg
06-15peak 64 on 07-1007-14
~43 posts a day across 16 accounts, about 2 to 3 per creator per day, seven days a week. The volume barely dips, even on weekends.
Posts by hourUS Eastern
12a6a12p6p
Almost nothing overnight. Posting ramps from late morning and clusters through the afternoon and evening. Hours are US Eastern, but the takeaway is the shape: post when your audience is winding down, not the exact hour.
Why it worked
Four things this campaign got right
7M views. The only trace of the app is a tiny “Created with Mozart AI” in the corner.
01
The hook is about the audience, not the app
The best video, "Crazy texts with my MIL" at 7M views, never mentions music. It is a mother-in-law demanding $45 for a Father's Day BBQ, unfolding text by text over footage of someone prepping dinner. The app appears only as a small "Created with Mozart AI" credit and as the song underneath. People came for the drama. The product rode in behind it.
02
One winning hook, then a saga
"Crazy texts with my MIL" was not a one-off. It became Part 2, a prequel, "before I give birth," "Part 4: texts between my husband and his mom." The dating angle "reasons hinge is traumatizing me" ran to 10 parts. Serializing a proven hook turns one hit into a returning audience that comes back for the next installment.
under 1K
1.1K
1K to 100K
199
100K to 1M
15
over 1M
posts by views, out of 1,2986
03
Volume covers for a low hit rate
The median post got 275 views. Most of the 1,298 posts went nowhere. That is fine, because the plan was never for every post to land. Sixteen creators posted 43 times a day so that a few drama hooks could catch, and then those winners got serialized and multiplied.
04
Mundane b-roll makes it feel real
No production, no talking to camera, no product screen. A kitchen, hands, chicken in a pan, and a text thread. That ordinariness is the point. It reads as a real person, which is exactly the trust a native feed rewards and a polished ad kills.
Five rules from this campaign
1
Lead with drama, not your product
In this campaign the product was literally the worst-performing hook. Open with your audience's problem, not your app.
2
Serialize your winners
When a hook hits, do not move on. Make a Part 2, a prequel, a "what happened next." Returning viewers are cheaper than new ones.
3
Post through a low hit rate
A median of 275 views is fine if the plan is volume plus a few serialized winners. Do not stop after a few flops.
4
Keep the footage ordinary
No production, no product screen. Over-produced kills the real-person trust that makes people watch to the end.
5
Rotate proven openers across accounts
A few drama openers spread across many accounts beats chasing a brand new idea every single day.
Free · the full hook file
Every hook from a 27.4M-view campaign, ranked
All 50 Mozart AI hooks with their real views, likes and format, as a spreadsheet you can sort and steal.
- The 50 hooks ranked by real views
- The exact winning openers, ready to copy
- A new viral app teardown in your inbox weekly
No spam. Unsubscribe anytime.
HookViewsLikes
Crazy texts with my MIL7.0M378K
Part 2 (prequel): Crazy texts from my MIL4.1M177K
Crazy texts with my MIL before I give birth1.7M80K
+ 47 more hooks
Questions, answered
How did Mozart AI grow so fast?
16 creator accounts posting about 43 times a day for a month, roughly 1,300 posts. Most flopped (median 275 views), but the plan was volume: post enough drama-hook variations that a few catch, then serialize those into sagas. Six posts crossed a million views and carried the campaign.
How does Mozart AI actually fit into these videos?
Mozart turns text into a song. Creators feed the drama in, get a track back, and it becomes the video's "original sound," credited on screen as "Created with Mozart AI." The product is genuinely used. It is just not the hook.
Why did some posts get millions and others get 2 views?
The 2-view posts advertised the app: they put song, aisong and aimusic in the hook. The million-view posts opened a drama the viewer had to finish. Same product, opposite result. The story is the hook; the product is the payoff.
Can DropUGC make ads like this for my app?
That is the point of studying these. DropUGC writes hooks in the winning style for your app, matches each to real creator footage, and drops finished ads every morning. Your first drop of 6 is free.
Want hooks like these for your app?
DropUGC writes them in the winning style, matches each to real creator footage, and drops finished ads every morning.
No card needed · 6 free creatives · yours to download
More teardowns
Figures are public engagement metrics (views, likes, comments) and observable posting activity captured from the campaign's public posts on July 18, 2026. Format mix is share of classified posts. Independent analysis of public content.