Giving a voice room a 3D face: Mass is on the Southeast Asia social top-grossing chart, the only one that "looks different"

A small Shanghai company's 3D avatar social app with 13.5M installs, ranked #14 on the Philippines social chart and #19 in Indonesia. There is nothing new in the gameplay: it takes the most profitable Chinese voice-room apps going global in Southeast Asia (gifts, guilds, leaderboards) and wraps them in a QQ-Show-style 3D avatar and outfit store. This piece covers how it got started, where the money comes from, where it stands now, and what a small team can take from it.

A 12.5M-view video, "Why you betry me!"

On 2023 Jul 21, Mass's official TikTok posted a 36-second video. The caption was just one line, "Why you betry me!" (misspelled), followed by two dozen hashtags: #avatar #metaverse #soulmate #romantic. The stars were two 3D figures acting out a scene of betrayal. By Nov 6, the video had 12.5M views and 78k likes, and it is still the most-watched post on the account (TikHub, 11-06).

Most comments are from Filipinos. The most-liked one says "THE FIRST POST HELP" (603 likes). Another Tagalog comment reads "Pov: Mga add", meaning "POV: those ads", the kind of ad you've already seen; others say "nag DOWNLOAD NA PO AKO": I've already downloaded it.

Three years on, the account's daily posts look different. In Nov 4 it posted "new two-person motorcycle, 11.11 into the blind box", with 196 views; in late October, player-submitted posts like "dating him" and "dancing together at Halloween", each with a few hundred views. In Oct 27, a Filipino player left a one-star Google Play review: "I've already spent over 60,000 diamonds on the pumpkin magic blind box, why still no special effect I want?"

Plot videos paid for acquisition bring people in, then blind boxes, outfits and social ties keep them and get the money. That is Mass over the last three years.

What it is

A 3D avatar social app developed by Shanghai Xuanji Network Technology (XJWL in the package name stands for "Xuanji Network"), backed by Shanghai Zhuzi Internet, which makes 3D anime games. Launched in 2023, it focuses on the Philippines, Indonesia, Thailand and Vietnam. The listing date on Google Play is 2026-09-03, which reflects a reset of the store page; it is not a new app.

The first step is character creation: five face styles (manga face, doll face, trendy face...), hairstyles, tattoos (catering to Thailand and Vietnam's tattoo culture), plus back accessories like angel wings and devil tails. Chinese overseas-expansion media Yangfan Chuhai compared it to an "overseas QQ Show". Then come homes, pets, fishing, mining and bug catching.

What really holds it up are the rooms: party rooms, game rooms, karaoke rooms, where 3D characters enter, talk on the mic and send gifts. The profile page shows gift records and "bindings", and the platform has wealth and gift levels, intimacy scores and guild leaderboards. The whole setup matches best-selling Chinese voice-chat apps in Southeast Asia like SUGO, Chamet and Poppo. The only difference is that the avatars are 3D people who can change outfits.

In one line: a 3D-avatar version of a voice room.

How big it is now

Read the two numbers together. First, top-grossing rank is far ahead of download rank: it is nearly invisible on the free chart but sits among a row of voice, live-streaming and dating apps on the grossing chart. That is the typical ledger of a voice room: a few big spenders buy gifts and pull blind boxes, and they support the whole product. Second, it has passed its fast-growth phase: in May 2025 it had 360k global downloads in a month (per Yangfan Chuhai's data), over 10,000 a day; now it's just over 2,000 a day (the two sources use different methods, so the decline is an estimate). The founder has said publicly "20M users", while Play shows 13.5M, and iOS is tiny. It is essentially an Android product.

Where users come from

Paid acquisition, mainly in the Philippines and Indonesia. Yangfan Chuhai's 2025 June teardown (citing Guangdadа data) is clear: the creatives are dramatic plot videos shot with 3D characters and realistic settings, the same format as the "Why you betry me!" opener. On the timeline, early traffic came from Thailand and the Philippines; in August 2024 it took off in Indonesia; in June 2025 Indonesia's single-day download peak exceeded 13k; in May 2025, the Philippines had 140k and Indonesia 100k, together 75.8% of global downloads (both per Diandian data).

Retention through community. The TikTok account has 146k followers and 1,232 videos, now mostly player-submitted content. Facebook groups, Discord and Instagram (21k followers) run contests with diamond rewards, such as the Halloween UGC contest at the end of October. These don't bring in much new traffic; they give existing players something to do.

Who plays. Founder Cheng Yinbin told Yangfan Chuhai in an April 2026 interview that the core users are young women aged 18 to 25, with men arriving through organic traffic. The comments back this up: one says they're introverted and no longer feel lonely after making friends here; a Thai player says "after playing I got a real girlfriend"; others warn "there's some dirty stuff posted here, be careful".

Where the money comes from

Blind boxes and outfits. Basic clothes cost purple diamonds, which you can earn through check-ins, tasks, and watching ads. Nicer high-tier outfits and player-made UGC works need blue diamonds, which are basically top-up only. The shop, the spin wheel, and limited-time blind boxes (11.11, that two-seat motorcycle is in a blind box) keep pushing people toward top-ups. One Indonesian review with 412 likes: "Fun to play, like Zepeto plus Roblox, but please make the clothes and props less expensive 😭".

Gift tipping and host revenue share. The official site massbloom.com says "receive gifts and get 60%, settled in real time," and the pitch is that with a 3D avatar you can "grow followers without showing your face." Southeast Asian voice-room host economies used to require showing your face, or at least your voice. Mass lowered the bar another notch.

UGC revenue share. Players design their own clothes and rooms, list them for sale, and creators keep 70%. The founder says creators in Korea, Japan, and Taiwan already earn around ¥10K a month from this (company claim).

Bypassing the platform cut. There's an Xsolla web store for direct top-ups, avoiding Google's 15–30% fee. A Steam version also launched at the end of 2025 (13 reviews, 46% positive, basically no traction).

Revenue isn't public, and the web top-up portion isn't visible from outside.

What players are complaining about

Going through Play reviews from late October to early November in the Philippines, Indonesia, and Thailand: lots of five-star ratings (many read like templates, likely partly from event rewards), with negative reviews concentrated on three issues:

• Too expensive, can't get what you want: "Over 60,000 diamonds and still can't get the effect I want"; "Almost every hairstyle and outfit costs money"; "Everything now requires expensive VIP."
• Lagging worse with each update: "Heavier with every update", "Phone gets hot", "Updates every time but I can't tell what changed"; one one-star Indonesian review says avatars in the lobby after an update "look ugly like clowns."
• People problems: abuse and harassment get reported and nobody acts; there's no way to find the account deletion option.

These are the old problems of voice rooms, and switching to a 3D skin didn't escape them.

Who it's competing with

The veteran 3D avatar social apps are all bigger than Mass, but none has deepened voice-room tipping; the voice-room apps all earn more than Mass, but none is 3D. In the top 30 social Top Grossing apps across the Philippines, Indonesia, and Thailand, Mass is the only 3D avatar product (Google Play charts, 11-06). I didn't find any new app clearly copying "3D avatar + voice rooms."

Where it's heading

Starting in April 2026, Mass has been going "all-in on AI": generating 3D clothes and rooms from a sentence or an image, creating a 3D avatar with about 90% likeness from a single photo, having an AI double play Werewolf, attend parties, and make friends for you, then telling you in the morning "how many friends you made and how many tokens you spent." Pricing is $20 / 30 / 50 / 100 for monthly token packs (founder interview). The Steam page mentions integrating hyper3d for AI generation.

It's a new story, but by November I see no lift in the numbers: about 2,000-plus new installs a day, and no sign that AI has brought in new users.

If it were you

C. Worth studying, not worth copying. Mass proves one thing: in a category already proven to make money, changing the skin can make you the only one that "looks different" on the charts, skipping the job of educating users. But this skin, meaning the 3D engine, the character creator, the huge catalog of outfits, and the UGC toolchain, took a game company backed by investors three years to build. Its volume is continually bought in the Philippines and Indonesia. Its money depends on guilds and big spenders. A small team can't get any of these three.

Who it suits: teams already running overseas voice-room apps with existing room, gift, and guild systems and Southeast Asia ad experience. Adding a virtual avatar layer to your own product (even a 2D or Live2D one) is much cheaper than building a Mass from scratch.

Who it doesn't suit: indie developers with a team of 1–3; people who haven't bought Southeast Asian ads or run guilds and hosts; anyone hoping to build traction from TikTok organic reach. Mass's official account now gets a few hundred views per video, which organic reach can't carry for this kind of product.

How to do it: Don't copy the 3D look. Copy the approach: pick a category where the top 30 Top Grossing apps all look alike (voice rooms, live streaming, dating), and switch to an appearance that users can tell apart at a glance. Use "no face shown" to lower the barrier on the supply side, paired with a high revenue share (it gives 60% on gifts and 70% on UGC). Use dual currencies: free currency from check-ins and ads, scarce currency only through top-ups. Open a web store to bypass store fees. If you only want the part Mass does poorly: the biggest complaint in reviews is "lagging and phone heating up after updates." A lightweight virtual-avatar party room that runs smoothly on low-end Android is a bet worth considering (inference).

How to acquire users: Follow Mass's playbook. Use avatars from the product to shoot dramatic short clips (betrayal, jealousy, chance encounters) and run them on Meta and TikTok in the Philippines and Indonesia. Use the same material for the official account and for user submission contests. Do the math first: this kind of product breaks even on a handful of big payers, and without guild and big-spender operations, paid acquisition easily loses money.

When: No need to rush for a time window. Mass has held this position for three years with no one following, which suggests the barrier is resources, not speed.

What would show I got it wrong: ① Mass's AI push really lifts new installs (back above 5,000 a day), meaning "AI doubles doing social" is a new opportunity worth revisiting; ② a second or third virtual-avatar voice app enters and holds its place in the top 30 Southeast Asian social Top Grossing, meaning the reskin path is easier than expected and the bar to follow is lower than I judged; ③ conversely, Mass drops out of the top 50 social Top Grossing in the Philippines and Indonesia, meaning this 3D skin can't hold up paid revenue, and there's no need to study it at all.