# The first ad from a songwriting app sells "finding the look that fits you": Tyvix and the app factory behind it

> Source: https://stringers.io/en/apps/play_aisong.generator.creatsong.music.html (Stringers / 什么值得做, generated 2026-11-08)
> Stringers answers one question every day: what small product can one or two people build in a few months and actually make money from. Numbers come from 15 databases we collect daily, each with source and date.
> For agents: these verdicts and the underlying databases (verified revenue, App Store / Google Play charts, ad libraries, GitHub, Product Hunt…) are queryable live over MCP — tools today / hot / products / directions / direction / data. Setup: https://github.com/komako-workshop/stringers (the user logs in at stringers.io → "配置 agent" to get a personal URL). Site index: https://stringers.io/llms.txt

> Developer DQL TechLab moved over the Meta paid acquisition playbook that worked on an AI video app. It started paid acquisition 12 days after launch, ran 744 ads in total, and bought installs in a category Suno had already educated the market on.

[Mobile App](https://stringers.io/en/apps/index.md) · Android · written 11-6

Tyvix - AI Song Generator · DQL TechLab

# The first ad from a songwriting app sells "finding the look that fits you": Tyvix and the app factory behind it

**What it is** An Android AI songwriting app: type a few lines, get a song, then drop it into a template to make an MV. Includes ads and in-app purchases.

**Why it took off** The developer, DQL TechLab, moved the Meta paid-acquisition playbook it had proven on an AI video app over to this one. Ads started running 12 days after launch, with 744 ads in total. In a category Suno had already educated the market about, it bought installs.

**C** **How you could follow it** Don't build AI songwriting; the category is already packed. Learn the horizontal copy: whenever a big company makes an AI category popular, launch a new brand name, keyword-based package names, and multilingual creatives, start paid spend within two weeks, and scale up whatever works.

Grade C: the category is a red ocean (Suno's original plus a dozen million-install clones). Tyvix has no product differentiation of its own, hasn't reached the top-grossing chart, and its in-app revenue is estimated at under $5K per month. It gained traction through the studio's existing paid-acquisition machine, which typical followers don't have; what you can learn is the playbook. Few people know about it yet

Data (updated daily) · first launch 2026-08-13 · 11-07 installs 622,870

Android · Total installs worldwide

Data 2026-10-29 → 2026-11-07 (10 days, 10 observations). Google's detail page only refreshes install counts every other day, so a flat day doesn't mean nobody installed.

Android · Ratings count (cumulative)

Data 2026-10-29 → 2026-11-07 (10 days, 10 observations)

Tyvix is an AI songwriting app on Google Play: type a few lines, get a song, then drop it into a template to make an MV. In under three months it has racked up 620k installs, but the product is no different from dozens of similar apps, and reviews include people complaining it doesn't use the lyrics they wrote. What drove it up wasn't the product but a ready-made Meta paid-acquisition machine at a Vietnamese app factory. Below: how it got going, whether it's making money, what you should learn from it, and why you shouldn't follow it into AI songwriting.

### An ad that got the wrong product

In 2026 Aug 13, Tyvix launched on Google Play. 12 days later, in Aug 25, it started running ads on Facebook. The earliest ads visible in the Meta Ad Library have the headline "👉 Commence maintenant ✨" (Start now), and the body reads:

> 💥 Découvre le look qui te correspond vraiment.

It means "discover the style that truly fits you," which is copy for AI outfit-change and AI portrait apps, not a songwriting app. More French copy of the same kind ran in Aug 26, Sep 11, and Oct 3; by Oct 27, the English version "Find the style that truly fits you" was still running. The first creative that actually talks about songwriting only appeared in Aug 31, in Spanish: "Crea canciones con AI en segundos" (Create songs with AI in seconds).

Running another app's copy word for word suggests these creatives were reused from the same developer's other apps. This is the key to understanding Tyvix: it's a product that came off an assembly line, not anyone's original work.

### How it grew

| Date | Google Play cumulative installs (AppstoreSpy estimate) | What happened |
|---|---|---|
| 09-03 | About 4,700 | Numbers start to show |
| 09-29 | About 76k | September: steady climb, a few thousand a day |
| 10-01 → 10-04 | 124k → 281k | about 160k more over 4 days |
| 10-27 | About 525k | 1–20k a day |
| 11-06 | About 623k | +98k in the last 7 days |

The jump in early October on day 4 most likely came from a budget increase or a new country going live (inference). After that, daily volume climbed from a few thousand to 10,000–20,000 and never fell back.

Other numbers all point to the same thing: the traffic was bought.

- **Only 784 ratings** (11-06) against 620k installs, about 0.13%. Apps with organic traffic usually sit well above that ratio. Paid users install, try it once, and don't review.

- **Only on the free chart, never the top-grossing chart** . 11-06 It appears on the free Music & Audio category chart in 8 countries: South Africa and Germany 95, Nigeria 96, Canada 100, the Netherlands 104, Mexico 139, the US 151, and the UK 191. It has never made any country's top-grossing chart.

- **88 live ads, but the page has only 412 likes** (Meta Ad Library, 11-04; including 744 stopped ads in total). A page with a few hundred likes running close to a hundred live ads is the typical look of a pure paid-acquisition page. 55% of creatives are video, the rest dynamic creatives; copy comes in five languages (English, French, Spanish, Portuguese, Arabic), served to Facebook, Instagram, and Audience Network, landing directly on the Play Store page.

Searching "Tyvix" on TikTok, X, Reddit, and in the media turns up nothing. It has no organic voice at all.

### What users get once they install it

The store page pitches it warmly: "A story, a feeling, a few lines you just typed. Tyvix turns it into a complete song in seconds." Three steps: write or paste text → pick a style → wait a few seconds for the song, then pick a video template to set to the beat. It says it suits "people who want a one-of-a-kind gift" and "content creators who need royalty-free music." At the bottom of the page it admits the MVs are made from "prefab templates."

User reaction tells a different story. The most-upvoted 1-star review on Google Play (12 people found it helpful):

> "Don't try this useless app. It gives you a place to write lyrics, but you don't need it at all. What it generates is their own childish stuff. Giving it half a star." — Daniel Mwangi, 10-09

Another one, 10-11: "I wrote lyrics with a slightly dark country feel, and what came out had nothing to do with what I wrote. At least it was pretty sad..."

Out of 4.0,784 ratings, 485 are five-star, 74 one-star, and 112 two-star. The complaints cluster around one thing: it says "turn your words into a song," but it doesn't actually put your words into the song.

### Where the money comes from: ads plus in-app purchases, but no visible payback yet

The listing says "contains ads" and "in-app purchases," with IAP items from $6.99 to $79.99. It has never charted in the Top Grossing list; third-party site AppGazers estimates its October IAP revenue at under $5,000 (estimate). Ad revenue is not visible from the outside.

An app running dozens to hundreds of paid ad campaigns a month while its IAP revenue sits in the thousands of dollars almost certainly loses money on the math (inference). It looks more like a studio "adding budget to see" while batch-testing, not a business that has already worked.

### Who's behind it: a factory copying the same playbook into every AI category

The developer on Play is DQL TechLab, and the developer entity registered on the store page is VINALED CORPORATION in Ho Chi Minh City, Vietnam. It has at least 6 apps under its name, with roughly 24.7M installs combined (Google Play, 11-06):

| App | Category | Total installs |
|---|---|---|
| Nexus - AI Video Generator | AI video | 12.4M |
| Lynix: AI Photo & Video Maker | AI photo | 9.06M |
| Cast to TV | Screen casting | 1.61M |
| EngLearn Speak: AI Tutor | AI English | 930k |
| Tyvix - AI Song Generator | AI song writing | 620k |
| Gravida: AI Photo Maker | AI photo | 100k |

Every app looks the same: a keyword-stuffed package name (Tyvix is aisong.generator.creatsong.music), a newly invented short brand name (Tyvix, Nexus, Lynix, Gravida), a privacy policy on Google Sites, a website URL pointing to an old project, and a support email with "aso" right in the name. Nexus has 1,100+ paid ads running on Meta (as of early November), and its first ad ran in Aug 18, one week before Tyvix's.

In other words: the AI video play worked, and the same machine was moved over to AI song writing. The product is replaceable; the machine is the asset.

### Who it's competing with

Almost everyone. When Tyvix entered, AI song writing on Google Play was already a red ocean (install counts are all 11-06):

| Product | Launched | Total installs | Notes |
|---|---|---|---|
| Suno | 2024-12 | 50.4M, +7k in the last 750 days | The original maker of the in-house model, and the category was educated by it |
| Donna AI Song & Music Maker | 2024-04 | 13.6M | A Turkish team, already on the Mexico Top Grossing chart |
| Music AI: MV Video Generator | 2024-04 | 7.35M | Sells the same "song + MV" |
| Rythmix / Jukebox / Mozart, etc. | 2024 | 4M–6.2M each | Similar text-to-song |
| Beatron | 2025-11 | 5.26M, +7k in the last 178 days | Reached 8× Tyvix's scale in one year and entered the Australian Top Grossing chart |
| OurSongs | 2026-05 | 966k | Specializes in "a song for a specific person" |
| Melozi | 2026-08-27 | 632k, +7k in the last 148 days | Launched two weeks after Tyvix, now at comparable scale and growing faster week over week |

"Song + MV template" is also offered by Sondo, Muvio, Shoom, and Yuna. Tyvix has no feature that no one else has. Even Melozi, which launched two weeks later with the same paid acquisition, is growing faster than it does. In this category, several identical machines are running at the same time.

### If it were you

**C** : Copy its playbook, not its category. What Tyvix proves is that one paid-acquisition machine can buy hundreds of thousands of installs in any AI category that a big company has already educated. It doesn't prove that AI song writing has room left, nor that it has made money itself.

**Who it's for** : small studios that already have one or two apps paying back through Meta or Google paid acquisition, and that have a creative team and a monetization template in hand. What you should learn is DQL's "horizontal copying": whenever a big company educates the market on an AI category, launch with a new brand name, a keyword-stuffed package name, and a multilingual creative set, start spending within two weeks, scale what works, and swap to the next one when it doesn't.

**Who it's not for** : indie devs who have never run ads and want to build an AI song-writing app on organic store traffic or social media. This category has the Suno original plus a dozen million-install wrappers. Tyvix's rating-to-install ratio suggests there is almost no organic traffic here. Revenue that can't be verified and apps that never chart are the norm on this path.

**How to do it** : If you must touch AI songs, don't build a general-purpose writing tool. Tyvix's biggest complaint is "not my words," and the most lucrative niche is "gift songs." But that business lives on the web with one-time payments, not on app subscriptions. If you build an app, change the category: find the next AI category a big company has just made popular, where Play still has only two or three wrappers, and attack it at Tyvix's pace.

**How to acquire users** : Meta paid acquisition, with creatives tested in multiple languages at once. Start by pushing installs and chart rankings in cheap countries like South Africa, Nigeria, and Mexico, then check whether high-price countries like Germany, Canada, and the US can recover the subscription cost. Creatives can be reused across your own apps to cut costs, but don't do what Tyvix did and run reskinned copy from a different app category on a song-writing app. That's wasting money (inference).

**When to get in before** : There's no deadline. The window for AI song writing has already closed after 2024 years. When the next category appears depends on big-company launch events.

**What would show this read is wrong** : ① Tyvix enters any country's Music & Audio Top Grossing chart in the next month or two, which would show wrapper song-writing can still pay back and the category is looser than it looks. ② Its live Meta ad count rises clearly from 88, while DQL launches no new category, which would show the factory has decided this line pays. ③ The reverse: if its live ad count falls to single digits by December and install growth stops, the factory has given up on it too, and even people who are good at buying traffic can't get in.

[Android store page ↗](https://play.google.com/store/apps/details?id=aisong.generator.creatsong.music)

**Research notes: competitor table, acquisition evidence, full research report**

Competitors

| Name | Where | Launched | Scale | How it differs from this one |
|---|---|---|---|---|
| Suno - AI Music & Songs Maker | Android | 2024-12-03(Play) | Play 50.4M installs, +7k in the last 750 days, #1 on the France Music free chart | The original model maker and category educator, also on iOS and web. Tyvix is a wrapper with no in-house model (unverified) |
| Donna AI Song & Music Maker | Android | 2024-04-26 | 13.6M installs, #91 on the Mexico Top Grossing chart | A similar wrapper that launched earlier, has charted Top Grossing, and has a working monetization model |
| Beatron - AI Music Maker | Android | 2025-11-28 | 5.26M installs, +7k in the last 178 days, #103 on the Australia Top Grossing chart | A similar new player that reached 8× Tyvix's scale in one year and is already on the Top Grossing chart |
| Music AI: MV Video Generator / Sondo / Muvio / Shoom / Yuna | Android | 2024-04 to 2026-04 | 600k–7.35M installs each | Also sells AI songs plus music videos, so Tyvix's MV templates are not unique |
| AI Song Generator - Jukebox / Rythmix / Mozart / MusicLab / Soniva | Android | 2024 | 2.4M–6.2M installs each | Similar text-to-song, launched one to two years earlier |
| Melozi: AI Song, Music Maker | Android | 2026-08-27 | 632k installs, +7k in the last 148 days | Launched two weeks later, comparable scale, growing faster week over week, same playbook in the same period |
| OurSongs: Custom Song Maker | Android | 2026-05-10 | 966k installs | Focuses on custom gift songs. Tyvix only mentions gifts in its description |
| Suno / Udio / Tad AI / MusicWave / aimusic.so web apps | Web | 2023–2026 | Not verified | Use the model directly on the web with a free quota. Tyvix is Android-only |

Where users come from

- **Verified** **Meta ads (Facebook/Instagram/Audience Network)** : metaads library: Page likes 412, 88 ads running right now (11-04), 744 total in history, earliest ad seen 2026-08-25 (12 days after launch); multilingual creatives in English/French/Arabic/Spanish/Portuguese, 55% video, landing directly on the Play package

- **Verified** **Developer portfolio / studio reusing paid acquisition capacity** : DQL TechLab's library has 6 apps totaling about 24.7M installs, and Nexus has 1,103 Meta ads running. Some Tyvix creative copy is from a dress-up app ("Find the style that truly fits you"), which shows creatives are reused across apps

- Inference **Store ASO / organic search** : keyword-stuffed package name aisong.generator.creatsong.music, developer email aso.dqltechlabstore@gmail.com. Play links in Google results carry organic UTMs; volume can't be estimated

- Inference **Google App Campaign and other paid channels** : Common for Play apps. We have no record of it in the Google ad library, unverified

- Inference **TikTok / X / media / Reddit** : TikHub, x-radar, and Jina search returned nothing relevant

## Why Tyvix - AI Song Generator (DQL TechLab, Google Play) could scale

> Package name aisong.generator.creatsong.music | First launch 2026-08-13 | Music & Audio | Contains ads + IAP $6.99–$79.99 | Research day 2026-11-06

### Bottom line

Tyvix is a "text in → AI-generated song → template MV" wrapper AI music app. The product itself is no different from dozens of other apps in the same niche. It reached about 3k installs in under 620 months by **running a mature app factory (DQL TechLab, with 6 apps in its library totaling about 24.7M installs) that copied the same Meta paid-acquisition playbook onto AI music, a category Suno had already educated** . It was driven by distribution tactics, not by the product.

### 1. Data: how it's growing

Source: googleplay library app_day / app_day_ext (AppstoreSpy estimates; single-day figures shouldn't be taken at face value), app_profile, 2026-11-06.

| Date | Total installs | Notes |
|---|---|---|
| 2026-09-03 | 4,662 | AppstoreSpy started tracking |
| 2026-09-29 | 76,022 | September climbed steadily, averaging 2–6K a day |
| 2026-10-01 → 10-04 | 124,393 → 281,466 | In 4 days it rose about 200K, clearly a single push of volume (more budget or new countries) |
| 2026-10-27 | 524,858 | Mid-to-late October averaged about 1–20K a day |
| 2026-11-06 | 622,870 | 7-day change +97,921 (app_profile) |

- Only 784 ratings (11-06) against 620K installs, **ratings per install about 0.13%** , far below what organic apps usually see. Classic paid-user pattern: installed, tried once, never rated.

- Rating 4.0, histogram [1-star 74, 2-star 112, 3-star 37, 4-star 74, 5-star 485] (app_snap 11-06). The most-upvoted negative review on the store page: "gives you a place to write lyrics but doesn't use the words I wrote, generates childish stuff" (Daniel Mwangi, 10-09,12 people found it helpful), and "wrote dark country lyrics, the output had nothing to do with them" (10-11).

- Charts: 11-06 has only appeared in the Music & Audio **free category chart** in 8 countries, ranking 95–191 (za/de/ng 95–96, ca 100, nl 104, mx 139, us 151, gb 191). It hasn't made any Top Grossing chart and isn't on the overall chart (data card).

- Revenue: no Top Grossing record. The AppGazers model estimates October IAP revenue under $5K, October downloads 69K–320K (our October measurement is about 400K, above its ceiling) [appgazers.com/reports/developers/android-DQL+TechLab]. The store listing says "contains ads", so monetization is a mix of ads and in-app purchases.

### 2. Competitors: who else does the same thing

An extremely crowded category. In our googleplay database, apps with AI song / AI music / song generator in the name or description number in the dozens (2026-11-06, app_profile):

| Product | Channel | Launched | Scale (Play cumulative installs / last 7 days) | How it differs from Tyvix |
|---|---|---|---|---|
| Suno - AI Music & Songs Maker | Android (also iOS and web) | 2024-12-03 | 50.4M / +750K; #1 on France's music free chart, 23 on South Africa's Top Grossing | The original model builder, the category educator; Tyvix-type apps most likely call third-party or Suno-like APIs (unverified) |
| Donna AI Song & Music Maker (MOBIVERSITE, Turkey) | Android | 2024-04-26 | 13.6M / +176K; #91 on Mexico's Top Grossing | An established wrapper, already on Top Grossing, monetization works |
| Music AI: MV Video Generator | Android | 2024-04-24 | 7.35M | Also pitches "song + MV" |
| Rythmix: AI Music & Song Maker | Android | 2024-10-29 | 6.18M | Peers |
| AI Song Generator - Jukebox (TWOWISE, Turkey) | Android | 2024-10-16 | 5.67M | Same category, already #157 on Nigeria's Top Grossing |
| AI Music & Song Generator(One Music Player) | Android | 2024-06-28 | 5.41M / +200k | Peers |
| Beatron - AI Music Maker | Android | 2025-11-28 | 5.26M / +178K; #103 on Australia's Top Grossing | A newcomer within the past year, already at 8× Tyvix's scale |
| AI Song Generator - Mozart(NineG) | Android | 2024-10-08 | 4.09M | Peers |
| MusicLab / Sondo / Soniva / Shoom / Muvio (two apps) / Yuna | Android | 2024-04 to 2026-04 | Each 600K–2.6M | Most also offer "AI music video / MV", so Tyvix's MV templates aren't unique |
| Melozi: AI Song, Music Maker(Mee Game) | Android | 2026-08-27 | 632k / +148k | Launched 2 weeks after Tyvix, already comparable in scale and growing faster week over week: same play, same period |
| MusicGPT - AI Song Generator | Android + web | 2026-02-06 | 597k | Has a web brand |
| OurSongs: Custom Song Maker | Android | 2026-05-10 | 966k | Pitches "custom songs to give someone" |
| Web: Suno, Udio, Tad AI, MusicWave, aimusic.so and others | Web | 2023–2026 | Not verified | Uses the models directly on the web, with a free allowance |

Bottom line: **there's no white space** . When Tyvix entered, a dozen-plus million-install apps already existed in this category on Play, and Suno itself has an official app.

### 3. Growth tactics: where users come from

#### 3.1 Meta paid acquisition (confirmed, main channel)

Source: metaads database advertiser / ad / advertiser_snap (collected from 2026-11-01 onward).

- Facebook Page "Tyvix - AI Song Generator" (page id 61593355811547, new ID range). **The page has only 412 likes but is running 88 ads right now (11-04; 11-01 is 84), with 744 total historical ads including paused ones** . Hundreds of likes with dozens to a hundred-plus ads running is the typical shape of a pure paid-acquisition Page.

- The earliest ad we see started **2026-08-25** , and they began spending 12 days after launch. Placements: Facebook, Instagram, Audience Network. About 55% are video; the rest are DCO (dynamic creative).

- The creatives are mass-produced in multiple languages: English "Create AI songs in seconds. 🎵", French "Crée des chansons avec l'IA en quelques clics", Arabic "حوّل فكرتك إلى أغنية مع الذكاء الاصطناعي", Spanish "Crea canciones con AI en segundos", Portuguese "Crie sua própria música com IA".

- A telling detail: several French and English creatives read "Découvre le look qui te correspond vraiment" / "Find the style that truly fits you". That's copy from an AI outfit or portrait app, moved directly onto a music app, which shows creatives and copy are reused across several apps from the same factory.

- Ad landing pages link straight to the Play package name (ad.play_id = aisong.generator.creatsong.music). Targeted countries cover English, French, Arabic, Spanish and Portuguese markets, matching the countries where it's charting (za, ng, de, ca, nl, mx, us, gb): cheap volume (South Africa, Nigeria, Mexico) plus some high-value markets.

- From 10-01 to 10-04, four days added about +200K installs, after which daily adds rose from thousands to 1–20K. The most likely explanation is more budget or more countries (inference, unverified).

#### 3.2 Developer portfolio and ASO (confirmed pattern)

- DQL TechLab has 6 apps in the database (AppBrain/AppGazers record 7, plus Homova - AI Home Design), totaling about 24.7M installs: Nexus - AI Video Generator 12.4M (Meta running 1,103 ads, first ad 2026-08-18), Lynix AI Photo & Video 9.06M, Cast to TV 1.61M, EngLearn Speak 930K, Tyvix 620K, Gravida 100K.

- The playbook is identical: keyword-stuffed package names (aisong.generator.creatsong.music, ai.photogenerator.aivideo.aivideogenerator.aiart), newly coined short brand names (Tyvix/Nexus/Lynix/Gravida/Homova), a Google Sites privacy policy, a developer website hosting an old project's web.app (aip456-viramind-ai-art.web.app), and a developer email aso.dqltechlabstore@gmail.com (the name literally says ASO).

- This is a studio that mass-copies the pipeline of "hot AI category → wrapper app → Meta paid acquisition → ads plus subscriptions to recoup." After AI video (Nexus) worked, the same machine was extended to AI songs in 2026-08.

#### 3.3 Other channels (checked, nothing found)

- TikTok: a TikHub search for "Tyvix" only returned Star Trek's "Tuvix" and unrelated videos. No creator content.

- X: x-radar shows 90 posts in 0 days.

- Press / Reddit / Product Hunt / HN: Jina searches return only store pages, AppBrain, APKPure and Facebook ad video pages. No coverage or discussion.

- Organic store search: Play links in Google results carry utm_term=tyvix+ai+song+generator (organic). There is brand search, but its volume can't be estimated. The keyword-stuffed package name shows they're doing ASO.

### 4. Why it took off

1. **The category is already educated** : Suno's official app has 50M installs and was still adding 7K over the last 750 days. The demand for "type a few words, get a song" already exists, so there's no market to educate. Tyvix only has to catch the long tail of paid users.

2. **The paid-acquisition machine came before the product** : the same developer's Nexus already has 1,100+ ads running and tens of millions of installs. Tyvix started spending 12 days after launch and has run 744 ads in about two and a half months (metaads).

3. **Volume from low-price countries** : among the charting countries, South Africa, Nigeria and Mexico rank best. That fits the common play of buying cheap installs, then recovering subscription revenue in high-value markets (us/gb/de/ca/nl).

4. **Ads plus in-app purchases** : the store lists "contains ads", with IAP ranging from $6.99 to $79.99. Most paid users don't pay, but ads recover a little from them. Revenue is still weak (no Top Grossing record, AppGazers estimates under $5K/month in IAP), and whether it pays back is **unverified** . It looks more like a "spend more and watch" project within a studio's batch testing.

### 5. Differentiation

- Features: text-to-song plus template-based music videos. MV templates from Sondo, Muvio, Music AI MV, Shoom and Yuna already exist, and "a song written for someone as a gift" is also taken by OurSongs. **There's no recognizable product difference** , and reviews even complain it doesn't generate from the user's own lyrics.

- The only real difference is distribution: a new brand name, keyword package names, multilingual creatives made in bulk, and paid-acquisition capacity shared across the studio.

- So differentiation is **not** why it worked. Paid acquisition is.

### 6. How many people have seen it

- **Few people know about it yet** : no discussion on X, TikTok, media or Reddit, and no public revenue. It only appears on data sites like AppBrain, AppGazers and SensorTower/Appfigures charts (Nexus showed up at #1 on SensorTower's Kuwait Art & Design chart).

- But the category itself is already crowded: AI song apps on Play alone number 10+ at the million-install level. Copying it makes no sense. Everyone is copying Suno.

- Same play, same period: Melozi (Mee Game, launched 2026-08-27, 632K, +7K over the last 148 days) and Tyvix launched almost together at similar scale, and Melozi is growing faster.

### 7. What to borrow

- The studio playbook: one set of paid-acquisition, creative and monetization templates, applied horizontally to every AI category that big players have just educated the market on (video → photo → home design → English → music). Each category gets a new brand-name app, and whichever one runs gets more budget.

- Start Meta ads within two weeks of launch, test multilingual creatives at the same time, and reuse creatives across apps to cut costs.

- ASO combo: package name keywords + short brand name.

### 8. Still unclear

- Is it paying back: no Top Grossing presence; AppGazers estimates IAP < $5K/month; ad revenue unknown. Whether the early-October spike came from more budget or another event is unverified.

- Underlying model: whether it calls third-party APIs like Suno/Udio or an open-source model is unverified.

- Whether they're also spending on Google UAC / TikTok beyond Meta (TikTok's ad library only shows Europe; not finding it doesn't mean it isn't running). Google App Campaign is likely another main channel for the Play app; unverified.

- Retention: the rating/install ratio is extremely low, and reviews complain about "lyrics I didn't write." Retention is probably poor; unverified.

### Sources

- googleplay library (agent-im server): app_profile, app_day, app_day_ext (AppstoreSpy estimates), app_snap, app, 2026-11-06

- metaads library: advertiser (Tyvix, Nexus), ad, advertiser_snap, 2026-11-01–11-04

- Google Play store page [play.google.com/store/apps/details?id=aisong.generator.creat…](https://play.google.com/store/apps/details?id=aisong.generator.creatsong.music)

- AppGazers developer report [appgazers.com/reports/developers/android-DQL+TechLab](https://appgazers.com/reports/developers/android-DQL+TechLab)

- AppBrain [www.appbrain.com/dev/DQL+TechLab/](https://www.appbrain.com/dev/DQL+TechLab/)

- Facebook ad video page [www.facebook.com/61593355811547/videos/](https://www.facebook.com/61593355811547/videos/)

- TikHub TikTok search for "Tyvix" and x-radar search for "Tyvix" (both returned nothing relevant), 2026-11-06

This article was written by AI from public data and the data we collect daily. Numbers are sourced, and inferences are labeled.