A backwards-singing trend ran for a year, and a Vietnamese utility built 4.48M installs by making the search term its name

Reverse Audio: Sing Challenge is an Android utility that plays recordings in reverse, made by the Vietnamese studio Mandala. At the end of October it was adding 20–30K installs a day, so it looks like a newly emerging app, but it has actually been on the store for a year, and there are more than twenty similar apps. It gets its volume not from the product but because its name is exactly the phrase people on TikTok keep telling each other.

A couple, a song played backwards

In Aug 2025 22, the US TikTok account @neeniesteenie posted a video just over three minutes long, titled "No Arnold Schwarzenegger doesn't sing but he's at the end". The game is simple: someone records a line of a song backwards, another person learns to sing it backwards, and then they play their backwards singing in reverse to see if it matches the original. That video later reached 26.2M views, and the second one, from Sep 3, reached 34.4M (TikTok, 11-06).

The reverse singing challenge blew up. To play it you need an app that can reverse audio, and the videos rarely name one. They just say "search reverse audio." A year later people are still saying it. A Brazilian couple's gaming channel, Canal Angil, posted a reversed-audio video almost every day from October 15 to early November, all with the same caption, "O nome do app é reverse audio" (the app's name is reverse audio). The October 18 video got 647k views.

The Vietnamese studio's app is called "Reverse Audio: Sing Challenge."

What it is

What it does is what the store page lists: record a clip, reverse it with one tap, adjust speed and pitch, then save it or share it with friends to take the challenge. No song library, no levels, no AI. The developer, MANDALA CONSULTING CO.,LTD, is a Vietnamese company. The same account also publishes a desktop pet, icon themes, pattern lock screens, a star map, a GBA emulator, and sticker makers: classic "utility + ads" apps.

Revenue comes from two places: ads, plus in-app purchases added in April 2026, at $6.99–$28.99. The store has 400 reviews. The three most-upvoted ones all say the same thing:

"Ads are insane, one or two on every screen, and a video or game ad between every step." "Too many ads, but I still kinda like it." "All ads. You can't do anything without watching a pile of them." (Google Play US store, 11-06)

It isn't new

The store lists the first release date as 2026-10-09, but this date was reset. The earliest version on APKPure, 1.0.2, was released on 2025-10-12, and third-party trackers log the listing date as 2025-10-06. By May 2026, cumulative installs were about 3.37M (AppstoreSpy estimate), with natural daily installs of 3,000 to 7,000 over those months. Roughly three-quarters of today's 4.48M came from before May. The same developer's desktop pet app also had its first release date reset in October. Why the dates were changed (account transfer, listing edit, or store expansion) is unverified.

The real change came in late October:

In November 6 it entered the Entertainment Top Free chart in 9 countries: #55 in Turkey, #80 in Indonesia, and between #100 and #200 in Malaysia, Saudi Arabia, France, Mexico, Argentina, Germany, and Sweden. It has never charted in Top Grossing.

It's hard to tell from outside who drove that surge. Two things line up in time: the Brazilian couple's channel, with 647k views on October 18; and from October 16 to 26, a cluster of TikTok accounts with Vietnamese region tags ("ReverseSingingforfun", "AI Reverse Audio") posting promo videos for the reverse challenge, captioned "One-tap reverse, singing challenge goes viral." These accounts never name the app, so they can't be attributed to Mandala. The developer also pushed two versions that month, which may have earned some extra store exposure (inference).

Why people install it

Several things stack up:

1. The name is the search term. The videos say "search reverse audio." Search that on Play and an app with the same title lands near the top. On iOS there's a same-named app from another Vietnamese team with 32k reviews. That reputation was built over there, and the Android listing rides on it.
2. The meme never died. From August 2025 until now, new people have posted every month. In October there were new challenge videos from the US, Italy, Finland, and Brazil, and the Italian one has 95k views.
3. Others pay for traffic, and it picks up the spillover. Plenty of creator videos on TikTok are tagged as ads, with captions like "search reverse singing on the App Store." They point to iOS equivalents, and there's no evidence Mandala ran them.
4. High volume, low quality. 4.48M installs produced only 428 ratings, about one in ten thousand; the top of the category runs at one in a hundred. Most people who install it delete it quickly. That ratio looks like traffic from ad networks or Google's app promotion. Whether it buys its own ads isn't visible in the ad libraries, unverified.

Who it's competing with

The feature is a single audio-processing function, and dozens of clones appeared within a year. On Google Play, more than 20 apps have reverse audio or reverse singing in the title, as of 11-06:

In this category it ranks fourth or fifth, a mid-tier volume play.

What's actually worth learning here

It isn't the reverse app. It's a playbook:

• On TikTok there's a pattern where creators tell viewers to "search this feature word" without naming the app. Whoever's name is that word gets the store search traffic.
• The meme outlived its peak and left a long tail. A year later it still gets several thousand organic installs a day, enough to support ad monetization.
• The cost is low enough to run in bulk: one feature, one studio website template, several developer accounts launched at once. Mandala and NindaFuit appear to work this way (inference).

If it were you

Verdict: The reverse-audio category is already taken: a leader at 12M installs and more than twenty clones. Launch now and you'll rank around #25. What you can take is the method: when the next "search this word" meme appears, race for the first week.

Who it suits: One or two people who can build and publish a single-feature Android app within three days, who already have ad monetization and multi-account experience, and who watch TikTok trends daily and react fast to new memes.

Who it doesn't suit: People who want a retained, branded product; people who only want iOS subscription revenue (these users come to play once); and people who can't handle bad reviews about "too many ads."

How to do it: Watch TikTok for challenge videos with tens of millions of views where commenters ask "what app is that?" Use the creator's exact wording as the app name (one English word, then localized store titles for the countries the meme spreads to: Portuguese, Turkish, Indonesian). Build only that one action, one tap, zero learning curve. Monetize with ads, plus a one-time purchase to remove them.

How to get users: Store search is the main channel: put the keyword in the title, the description, and the first screenshot. Run one or two TikTok accounts too, filming challenge videos with your app and saying outright "the app is called xxx." When the meme spreads to a new country (Brazil, Turkey, and Indonesia this time), do another round of store copy and videos in the local language.

When to move: One to two weeks after the meme appears. This reverse challenge peaked in August 22, and in September a wave of similar apps launched. The leader, AI Tech Expe, didn't launch until November and still reached #1, so don't wait more than three months, and buy more volume than the others.

Signs I've misread it: ① The next word-category meme comes out, and apps launched in week one hit fewer than 100k installs within a month. That means store search is already saturated by big accounts running paid acquisition, and this path is closed. ② The last 7 days of daily growth for the head of the reverse-play category drop to under half their level before December. That means the long tail is shorter than I thought, and "a dying meme can still feed you for a year" doesn't hold. ③ Google Play starts delisting or demoting batches of same-name, same-template spam apps. Then the multi-account route is closed.

Data and sources

• Installs, ratings, charts: Google Play store pages and install counts (10-29 to 11-06); daily-growth early figures are AppstoreSpy estimates; last 30-day growth comes from appdashboard.app.
• Launch dates: APKPure version history (earliest 1.0.2,2025-10-12), AppstoreSpy (2025-10-06).
• Reviews: Google Play US store page, 11-06.
• TikTok: @neeniesteenie (2025-08-22, 2025-09-03), @canal.angil (2026-10-15 to 11-02), Vietnam-region reverse promo accounts (2026-10-16 to 10-31); view counts as seen on 11-06.
• Competitors: Google Play app pages and install counts (11-06); the same-named iOS app from the App Store (id6754755473).
• Anywhere marked "inference," "estimate," or "unverified" is inference, estimation, or unconfirmed.