AI Couple: a "couple hug video" app hits the Italian category chart, but it's really a template split out of a bigger app

AI Couple: Video Generator lets you upload two photos of a couple and generates short videos of hugs, dances, dates, and weddings. In early November it added 9k installs in a week and reached #12 on Italy's free Art & Design chart. But it isn't something a new team built from scratch: a Hong Kong studio split the "couple" template out of its flagship app, which has 550k installs, listed it separately, and then paid to push it up a low-bar chart. Below: how it got here, where the money comes from, how crowded this path already is, and what you can learn from it.

A one-star review: "4000 credits vanished"

On Oct 31, a US user left AI Couple a one-star review on Google Play: "Bought it, and the free credits the ads promised never showed up. I reported it and nobody replied. 4000 multiple credits missing. False advertising, I'm going after this!!!!"

A one-star review from a week earlier: "Any decent AI video app lets you watch ads for coins. This one doesn't have a single one. Don't want to pay? Then don't play." An Italian user just wrote "paid only", and a French user wrote "Bought it, still tons of limits." The most polite one, from Nov 3, was three stars: "It's okay, I just wish it were free and had ads."

These reviews actually make the business clear: no free trial, a paywall as soon as you open it, selling credits and subscriptions. The in-app prices on Play run $7.99–$49.99, and in the Italian store €8.49–€54.99 (store page, 11-06). It's built to collect money first.

What it is

A template-based AI video generator. Pick a template (beach sunset date, city night walk, wedding, Valentine's surprise, anime-style couple), upload photos of you and your partner, and a few dozen seconds later you get a video you can post straight to TikTok, Reels, or WhatsApp. It also bundles generic features: image-to-video, text-to-video, video extension, and text-to-image.

The store page is blunt about it, naming Seedance 2.5, Kling 3.0, and Seedream 5.0, the video model APIs from ByteDance and Kuaishou. It has no models of its own. What it builds is the templates, the interface, and the paywall.

How fast it grew, and how small it is

It launched Sep 23 and only collected 3,500 installs in its first five weeks. It sped up suddenly at the end of October, roughly 1,700–1,900 installs a day, gaining 9k in one week, and now it has just crossed the "10k+" tier.

11-06 rankings: #12 on Italy's free Art & Design chart, #19 in France and Sweden, #20 in the Netherlands, #28 in the UK, #31 in Germany; #21 on Germany's top-grossing chart in the same category. The best US free-chart position in the same category is only #138.

These ranks look good, but the category's bar is very low in small European countries: on the same day, the #15 app on Italy's chart had only 99k installs total. A thousand or two installs a day is enough to land in the top 15 there. Also, in more than a dozen countries (India, Nigeria, Turkey, Saudi Arabia, Indonesia, Mexico…) it ranks #54–59 on the top-grossing charts with nearly identical positions, which suggests these small markets share one chart and shouldn't be read as people paying in every country. The only revenue signal with real weight is #21 on Germany's top-grossing chart. Revenue estimate is small (inferred).

4.2,87 ratings. The distribution has only 1-star 17 and 5-star 69, with none at 2–4 stars, which looks like the app is only sending happy users to rate it on the store. But public text reviews include three-star and two-star ones, so the distribution chart may be out of date. We can't tell for sure.

Who's behind it

The developer is MOON TEMPLE LIMITED, registered in Hong Kong (address in Kowloon City). The website is a traditional Chinese company intro, the support email is Gmail, and the privacy policy is hosted on a free Google Sites page. The "website" link on the store page is admob.moontemplelimited.com, with a subdomain literally called admob.

It has 6 apps on Play, but the real weight is VifunAI : AI Video Generator: launched 2026 February, 550k installs total, 60k gained in a week, #4 on the Germany and UK Art & Design top-grossing charts, #8 in the US, in-app prices from $2.99 to $179.99. VifunAI's description already includes "create couple moments." AI Couple takes that template category and lists it on its own; the same company also split out an AI Dance, which hasn't taken off yet.

So the plain version is: the flagship app proved AI video templates make money in Europe, and the studio carved out the most emotional, most giftable "couple" template and ran it again under a narrower name. The backend, model APIs, credit system, and paywall are all ready-made, so adding one more package costs almost nothing.

Where users come from

The publicly visible social buzz is zero: in nearly 60 days, no X post mentions AI Couple or VifunAI, and no official accounts or creators on TikTok promote it. In that case, a steady gain of over a thousand a day across 29 countries most likely comes from:

1. Paid acquisition (inferred). Google App campaigns or Meta. Public ad platform records don't show this company, but that doesn't mean it didn't spend. The sister app VifunAI sits near the top of the grossing charts, which suggests they have an acquisition model that pays back.
2. Cross-promotion from the flagship (inferred). VifunAI brings in 60k new users a week, and promoting "an app dedicated to couple videos" inside it costs close to nothing.
3. Keywords (visible). The title is literally "AI Couple" and "Video Generator," the description is stuffed with hug, dance, wedding, anniversary, Valentine, and long-distance, and model names like Seedance and Kling are written in too.
4. Small-country category charts. Bought traffic pushes it into the top 20 of the categories in Italy, France, and the Netherlands, and once it's charting it picks up some organic downloads.

How crowded this path already is

Very crowded, and the competitors are bigger:

The AI hug video trend started at the end of 2024 with Vidu and CapCut templates. On TikTok, @guma_cook's tutorial from November 2024 has 2.48M views. AI Couple arrived a year and a half late and is the smallest in its category.

The demand is still there. TikTok still has new "AI couple hug" videos from the past 30 days. In November 4, @capisean_capisara posted an AI couple video captioned "send this to your love," tagged "made with Higgsfield AI," with 670k views. In late October there were also tutorials in Burmese and Sinhala on "making a couple video from two ordinary photos." Users want it, but there are now so many tools for it that they can pick any one.

Where it falls short

• No free first run. Most competing apps make users watch ads for credits, or give a free watermarked clip. This one asks for money right away, and nearly all the bad reviews are about exactly that.
• Customer support. "Credits never arrived, nobody replied" is the kind of bad review that hurts a paid product most.
• No differentiation. The models belong to someone else, the templates are close to a dozen similar apps, and the only difference is a name and keywords that lean more "couple."

If it were you

In one line: C. The playbook is worth learning, but copying an AI couple video app isn't worth it. The category already has a dozen million-download rivals, the model makers themselves give away free templates, and this app's own volume and revenue are both small.

Who it suits: a team that already has a paying AI image or video app, with a working backend and a paid acquisition model that pays back. For you, the move to learn is "pick the best-converting template out of your main app and list it separately," which costs almost nothing.

Who it doesn't suit: solo indie devs starting from zero with no paid acquisition experience. This category lives on paid ads and keywords, not organic spread, so you'd build it and nobody would see it. Model APIs charge per use, so if you don't collect money you lose money, and if you do, you get the same bad reviews this one got.

How to do it: don't make "yet another couple video." Pick an emotional scene nobody has over-built: hugging your younger self, a video with a loved one who has passed, a memorial video for a pet. The growth of Hug My Younger Self and Rendio suggests a different scene makes a big difference. Or change language or market: people on TikTok in South and Southeast Asia teach how to make couple videos in local languages, which suggests demand there, but there are few localized apps (inference). On the product side, fix what this one gets wrong: the first clip free with a watermark, ad-for-credits, and someone who actually handles credit delivery.

How to get users: start by posting finished clips to TikTok and Reels, with captions like "send this to your love" tied to a specific holiday or anniversary. Then run Google App ads with scene keywords stacked in the titles and descriptions. The month before Valentine's Day (Feb 14) is the best paid window of the year.

Deadline: if you want to hit Valentine's Day, launch and prove paid conversion before mid-January.

Signs you've read it wrong: (1) If AI Couple's daily installs fall back to a few hundred within two or three weeks, even a studio with paid acquisition experience can't make this slice pay back, so don't follow it. (2) Conversely, if it climbs into the top 10 of the Top Grossing chart for Art & Design in Germany and the UK in December and January, the "couple" slice converts better than estimated and a better version is worth building seriously.