Original · Unique Research · 2026-06-16
Editor's note: The first-person report and its judgments belong to the original Chinese author. This English rendition retains the May 2026 global AI Web data narrative (head landscape, mid-tier true-growth vs. statistical illusion, China-market variables, three trends, closing), plus the publisher's list-detail, methodology, concept definitions, metric definitions, and free-use disclaimer. All figures, company and product names are preserved. Market and product figures are source attributions, not independently verified findings.
UNIQUE RESEARCH
Global AI Web Data, May
Numbers talk, but numbers can also "lie"
If you've been spammed by Claude Fable 5 these past two days, don't get excited just yet.
On June 9 Anthropic dropped this so-called "Mythos-grade" new model — always-on adaptive thinking, 128K output, a 1-million-token context window. Sounds intimidating, right? But you may not have noticed that on May 28 Claude Opus 4.8 had just been released, only 42 days apart. This isn't iteration anymore; this is Anthropic telling the whole world: "we're in a hurry too."
Hermes grew 56% in monthly active users in May, and traffic rose 46%. Its trait is going open-source, making tool-calling highly precise, with a strong developer-community reputation, and it's one of the few claw-type Agent apps with real growth. In fact, technically literate users are migrating toward products with genuine capability, while marketing-driven traffic is bleeding out. On the domestic side, skillhub (a skill platform) +6.5%, nanoclaw +24.8%; these small products bucked the trend with small May gains, following similar logic — users are filtering for genuinely useful small tools rather than crowding the top products.
May's global AI Web product data is out now; some results are as expected, some completely unexpected. Let's look together.
The Head Landscape: ChatGPT Still Growing, but Gemini Is Chasing Harder
First the top few.
ChatGPT: May monthly active users 510 million, up 3.14% month over month.
Gemini: May monthly active users 335 million, up 8.8% month over month.
Claude: May monthly active users 101 million, up 12.78% month over month.
On the surface, ChatGPT's scale is unbeatable — 510 million MAU, nearly 200 million more than second-place Gemini. But growth rate tells another story: Gemini's growth (8.8%) is almost triple ChatGPT's (3.14%). Claude is more striking, at 12.78% growth; though its absolute base is still only one-fifth of ChatGPT's, if this speed holds...
Wait, don't rush to say "Claude is making a comeback." In May Anthropic did advance a wild plan: signing to purchase Google's next-generation TPU compute, contract effective from 2027, 200 billion US dollars over five years at 5 gigawatts, with Google TPU pricing 40% to 50% lower than Nvidia's solution. At the same time it took over 220,000 GPUs in SpaceX's Colossus 1 data center. The first thing after compute arrives is to unshackle users: Claude Code limits doubled, and Opus API rates were greatly relaxed.
Worth noting: the AI index released by Ramp shows Anthropic's enterprise paid adoption (34.4%) surpassed OpenAI's (32.3%) for the first time in April. A year ago this figure was only 9%. Claude Code contributes 4% of public GitHub commits worldwide.
At May's Google I/O, Google went all-in on "Agentic AI" — Gemini 3.5 Flash, Gemini Spark as an always-on personal assistant, and Gemini Omni that can generate video. This time Google wasn't launching products, but sending a signal: the search giant finally found its comfort zone in the AI era — covering your whole day. Daily Brief emails you a schedule summary in the morning; Gemini Live integrates into various apps; Google Health and Fitbit Air are connected in too.
This explains why Gemini's growth is climbing: it's no longer just a chatbot, it's becoming the operating system inside your phone.
But after I/O, a controversy trended. Google changed Gemini's billing rule from "daily prompt count" to "compute-based billing"; paid users frequently hit caps during intensive coding and multimedia tasks, hitting 5-hour cooldowns and weekly quota locks. Complaints exploded on social media, and Google urgently raised quotas 9x to calm things down. This episode shows Gemini's user activity is indeed rising so fast that Google's billing infrastructure couldn't keep up.
And ChatGPT? A 3.14% growth rate isn't low for a product with 500 million users. On May 5, GPT-5.5 Instant took over as ChatGPT's default model, reducing hallucinations in high-risk scenarios by 52.5%, giving more concise replies, and adding cross-conversation memory and Gmail personalization. This upgrade was quiet for existing users but helps retention.
But what OpenAI truly pushed in May wasn't the chatbot, but Codex.
Codex's weekly active users surged from 3 million in early April to 5 million in early June, up 66% in two months. More crucially, the user mix is changing. 20% of users are no longer programmers, but knowledge workers making reports, writing slides, and running data analysis. The Chrome extension launched May 7, mobile launched May 15, Windows computer-control opened May 29, and Goal mode went live letting Codex autonomously execute tasks for hours. OpenAI reset Codex usage limits once on May 17 and again on June 2; Altman's promise is to reset it every time it crosses 1 million users. Codex has now gone from "a programmer's coding assistant" to "a work agent for everyone," and is about to be integrated into the main ChatGPT interface.
On May 16, OpenAI also did something big: a major reorganization. Co-founder Brockman unified the ChatGPT, Codex, and API product lines; the video-generation product Sora was shut down the same day, and product lead Kevin Weil and technical lead Bill Peebles left. The 1-billion-dollar Disney collaboration fell apart because compute consumption badly mismatched revenue: Sora burned 15 million of compute a day for a lifetime revenue of only 2.1 million, with a 1% 30-day retention. This decision was cold-blooded, but also correct. Video generation looks cool but makes no money; Codex survived because knowledge workers will pay.
In plain terms: OpenAI is going from a "technology company" to a "business company." Q1 revenue was 5.7 billion US dollars, leading Anthropic by about 1 billion, but adjusted profit margin was -122% — losing 1.22 dollars for every dollar earned. The IPO is in sprint, with valuation expectations above a trillion dollars, but the profit path hasn't run through.
Mid-Tier Shadow War: First See Clearly Who Is Truly Growing, and Who Is a Statistical Illusion
OK, the giants' side every media is reporting. Let's look at the more interesting part — the mid tier.
But before talking about the mid tier, I must pour a bucket of cold water: the numbers on the growth list aren't all real.
First, how Pixelcut's 1200% growth came about
On the growth list Pixelcut shows 6.65 million MAU, up 1215% month over month — 12x. I was shocked when I saw this number; my instinct reaction was "the image-editing track is about to explode."
But I searched the background and found it's nothing of the sort.
On March 3, 2026, Pixelcut officially announced a brand rename to Pixa, migrating the web entry from pixelcut.ai to pixa.com. At the time it said it served over 15 million users monthly. But after the rename they found a problem: old users searching "Pixelcut" couldn't find it, and new users didn't know who "Pixa" was. A few months later, the official quietly changed the brand back to Pixelcut.
In other words, March-April traffic was scattered to pixa.com, and after reverting to pixelcut.ai in May, the brand-search and tool-search traffic originally belonging to Pixelcut flowed back. The "1200% growth" in the data table is essentially a statistical-basis correction after a domain migration, not a product blow-up.
Pixelcut's largest May traffic source was organic search, at 53.52%, with core keywords being brand and tool words like "pixelcut," "remove background," "background blur." Paid search was only 1.41%. This wasn't ad-spend driven; it was old users coming back to find the product.
The true-growth group: these products' rises have solid business support
Wanderboat — MAU 5.51 million, up 65% month over month.
Early Wanderboat solved "make me a travel itinerary"; this need looks big but is actually low-frequency. Now its app-store expression has clearly shifted to "AI Local Guide," focusing on restaurants, cafes, nightlife, weekend activities, date spots, and nearby hidden gems — essentially turning a Trip Planner into an "AI version of Google Maps + Yelp + local-life Xiaohongshu." This shift is key because users don't necessarily travel every month, but every week they may need to find a place to eat, meet friends, date, or check in. In May it also caught North American summer travel and local-event season, plus mobile updates, creator seeding, and the spread of visual recommendation cards, so growth was easy to amplify. Wanderboat's signal value is that an AI-app opportunity isn't necessarily reinventing a big tool, but transforming a low-frequency scenario into a high-frequency entry: from "help me plan a trip" to "where do I go today."
candy.ai — MAU 28.84 million, up 24.03% month over month.
Who is this? An overseas emotional-companionship product. While Character.AI (MAU 7.56 million, down 11.79%) and SpicyChat.AI (MAU 12.21 million, down 13.37%) both fell, candy.ai bucked the trend with a 24% surge.
candy.ai positions itself as a "visual-first AI companionship platform" — emphasizing deep character customization, appearance adjustment, personality settings, plus a February video upgrade and a real-person mode. 50 million registered users, 87% retention. This data is backed by user base and retention, not hollow.
The emotional-companionship track is undergoing a brutal reshuffle. Character.AI's former star halo is fading — aging product experience, hard monetization, user aesthetic fatigue. SpicyChat is also falling. Meanwhile lighter, more flexible products like candy.ai and JanitorAI (up 8.22%) are nibbling the gap left behind.
This track's logic looks like early social apps — what users want isn't the "smartest" AI, but the AI that "understands me best."
Baidu AI Search — MAU 14.09 million, up 22.34% month over month.
This data has clear business-logic support. On May 9 Baidu released Wenxin 5.1, with the core technology "multi-dimensional elastic pre-training" — total parameters compressed to about 1/3, activated parameters to about 1/2, with pre-training cost only 6% of industry models of the same scale. LMArena search ranking first domestically and fourth globally; AIME26 math reasoning 99.6 points.
Wenxin 5.1's capability was directly infused into Baidu AI Search, with the smart box upgraded, DeepSearch, and multimodal input rolled out across the board. At the same time Chengpian (writing software) rose 32.89% and Wenxin Yiyan rose 14.53%. Three products rising at once shows it's no accident — Baidu's "AI engine" transformation strategy is taking effect, now backed by underlying-model technology.
Meshy (4.78M MAU, +46%) and Tripo (3.08M MAU, +80%) — the two 3D-modeling champions.
These two going-global products' growth map to very concrete business events.
On May 21 Meshy partnered with FlashForge to launch an integrated "image-to-3D-print in one click" — calling Meshy's AI 3D generation directly inside FlashForge's slicing software, auto-mapping textures to filament colors. Meshy claims a 97% slicing pass rate. This is a big deal for the 3D-printing industry.
Tripo is even stronger — it has raised 50 million US dollars from Alibaba and Baidu Ventures, with 6.5 million+ creators, 90,000 developers, and nearly 100 million cumulative 3D models generated.
In May, AI 3D quietly exploded in a corner most people overlooked. Text-to-3D and Image-to-3D technology finally evolved in 2026 from "demo-usable" to "production-usable" — Meshy's model can be imported directly into slicing software to print; that's the key turning point.
DeepSeek Open Platform — MAU 6.5 million, up 47.93% month over month.
DeepSeek itself (the chatbot) has MAU 39.99 million, up 6.89%, ranking first domestically. But the Open Platform's 48% growth is more notable — it shows developers are pouring into DeepSeek's ecosystem.
In May DeepSeek completed its first external financing, targeting up to 50 billion RMB, with founder Liang Wenfeng personally contributing about 20 billion (40%), at a post-money valuation of 350 billion RMB (about 50 billion US dollars). More crucial, V4.1 is scheduled for June, adding MCP protocol support and image/audio multimodality. The MCP protocol is the "universal language" for inter-Agent communication; DeepSeek supporting it first means it's laying out the Agent ecosystem, not just selling APIs. Money is in place and the technical roadmap is clear, so developers naturally follow.
Ghost Hand Editing (鬼手剪辑) — MAU grew from 40,000 to 80,000, nearly 100% month-over-month growth.
This is a domestic going-global AI video-editing plus dubbing tool, targeting the short-drama going-global market. Short-drama going global is one of the biggest content-industry trends of 2025-2026 — Chinese short dramas translated into English, Spanish, Arabic, with AI dubbing replacing the original voiceover, at 6x lower cost than human dubbing, finishing a drama in half an hour.
Ghost Hand's growth isn't from any technical breakthrough in the product itself, but from stepping into an exploding external-demand market. This kind of "tool follows the industry" growth is often more solid than "technology-driven" growth.
OiiOii — MAU grew from 360,000 to 570,000, +60% month over month.
A domestic AI animation tool where 7 AI agents collaborate (art director, screenwriter, character designer, storyboard artist, etc.), turning one sentence into a narrative animation of 60+ seconds. 150+ animation styles, ramping up via invite codes in beta.
It landed exactly at the May turning point when AI video shifted from "single-clip generation" to "animation production flow." OiiOii isn't an ordinary text-to-video tool; it packages itself as an "AI Animation Agent Team": script, character design, storyboard, scenes, video generation, sound, and other steps are strung together by multiple agents, better suited for anime shorts, character IP, continuous storylines, and social-media short videos. Its new-features page also pushes this direction: one Canvas integrating 25+ image/video models, 140+ styles, batch concurrent generation, associative editing, first/last-frame and reference-image control — essentially selling an "animation workstation," not single video generation. This is attractive to creators because after video models like Seedance, Kling, Sora, and Vidu densely entered content-production discussions in May, everyone's question was no longer "can it generate a cool shot," but "can it stably produce a series of content with consistent characters, consistent style, ready to post on TikTok / Reels / YouTube Shorts." OiiOii also took clear growth actions: website redesign, new features, launching Oii TV to showcase user work; platform campaigns require users to post with the OiiOii watermark and #OiiOii tag to X, Instagram, YouTube, TikTok, Facebook, LinkedIn for a lucky-draw chance, which naturally amplifies UGC spillover. So this growth isn't simply "AI video is hot again," but anime/short-drama/character-IP creators starting to look for tools closer to a production line. OiiOii is capturing the dividend of "after AI video-model capability improves, creators need an upper-layer workflow to orchestrate the models." Its current risk is also clear: a multi-agent workflow sounds high-end, but the learning cost and generation stability may still scare off light users; if retention doesn't follow, this +60% may look more like a traffic spike from the new-product and campaign period.
Several New Variables in the Domestic Battlefield
In May the domestic AI-product landscape showed some notable signals.
DeepSeek held onto first place in domestic MAU (39.99 million), with 6.89% growth not stunning but healthy given its scale. More notable is the Open Platform +48% — the developer ecosystem is the moat.
Baidu's "three arrows together" — AI Search +22%, Chengpian +33%, Wenxin Yiyan +3.45%. Baidu is going from "search engine" to "AI engine"; this process is painful, but data is starting to give positive feedback.
ByteDance's "Coze" — visit growth 20.84%, leading the domestic growth list. In May ByteDance clearly made agents a strategic focus.
Meanwhile, Doubao on May 4 disclosed three paid tiers for the first time: Standard 68 RMB/month, Enhanced 200 RMB, Pro 500 RMB. The base version stays free. Doubao's April DAU already broke 140 million, MAU 345 million, firmly first among domestic AI assistants — scale is large enough that starting to monetize is a reasonable choice.
By contrast, Doubao's growth is only 2% (tied to experience volatility during the monetization test), Jimeng fell 10%, and Volcano Ark fell 14% — ByteDance's resources are shifting from "general large models" toward "agent platforms" and "paid conversion."
Alibaba's "Quark" — MAU 19.63 million, up 7.73%. At May's Alibaba Cloud Summit, Alibaba released Qwen3.7-Max and self-developed chip "Zhenwu M890," formally completing its full-stack AI system. As the flagship AI-search product, Quark grew steadily.
Industry Trends: Three Big Trends Behind the May Data
After these data, let's discuss a few trend judgments.
Trend one: general chatbots enter a "plateau," while vertical-scenario AI starts to explode.
The growth divergence among the three giants ChatGPT, Gemini, and Claude is obvious — ChatGPT moved from "high growth" to "steady growth," while Gemini and Claude are still eating market share. Meanwhile, emotional companionship (candy.ai), 3D modeling (Meshy/Tripo), AI search (Baidu AI Search), AI animation (OiiOii), and short-drama going-global tools (Ghost Hand Editing) all posted solid growth.
This looks a lot like mobile-internet's trajectory — first one super app rules everything, then vertical apps come out to slice up scenarios. AI is walking this path too.
Trend two: the emotional-companionship track is undergoing a brutal reshuffle.
Character.AI went from star product to -12% MAU in just a few months. This track's user loyalty is lower than imagined — wherever the experience is better, the characters richer, the interaction more natural, users run there. candy.ai up 24% while Character.AI down 12% is essentially a user migration.
Trend three: Chinese AI is finding its own rhythm.
DeepSeek raising 50 billion, Baidu AI Search +22%, Coze leading growth — domestic AI products are no longer "following America," but have found their own playbooks in search, agents, and open platforms. Alibaba Cloud's full-stack AI system (chip + cloud + model + inference) also shows China's AI competition is upgrading from the "model layer" to the "system layer."
Closing
In early June Anthropic released Claude Fable 5, and the whole industry was again drawn into the new round of model-capability race.
But if you look closely at May's data, a deeper story emerges: the AI battlefield is shifting from "whose model is smarter" to "whose product understands users better." The ceiling of general chatbots is faintly visible, while the mid-tier products truly solving concrete problems — whether emotional companionship, 3D modeling, or short-drama going-global tools — are finding their own growth room.
candy.ai's users don't care whether Claude Fable 5 has 128K output. Meshy's users don't care whether Gemini Omni can generate video. Ghost Hand Editing's users don't even follow the competition among the domestic big-five models. They only care about one thing: can this AI help me solve today's problem.
Sometimes the biggest trend hides where you didn't notice — but sometimes, the 1200% growth on the data table is just a brand changing its name.
Unique Research - authoritative global AI market insight organization
The detailed list of Unique Research's May 2026 AI Web ranking is as follows:
AI website-traffic list: TOP100 list of AI Web products' visits and monthly active users, split into global and China markets.
All sorted lists also provide growth lists, helping investment institutions discover promising projects, and founders and developers discover potential opportunities.
Our influence and recognition:
Unique Research's data and insights are widely trusted and cited by top domestic and international investment institutions, academic units, and media, becoming an important benchmark for market analysis, investment decisions, and trend judgment. Our partners and citers include (but are not limited to):
Securities firms and investment banks: Sinolink Securities, China Securities, Huayuan Securities, Haitong Securities, etc.
Authoritative media: China Securities Journal, TMTPost, Jiemian News, etc.
We firmly believe only data that survives verification by industry core participants has lasting value.
We believe in the power of open source and openness; Unique Research hereby makes three commitments:
This AI product list is permanently open-source and freely available; the complete methodology and raw data support full reproduction.
This list's data sources are strictly limited to independent third-party monitoring platforms; data self-declared by enterprises is not accepted as an evaluation basis.
This list's ranking system remains absolutely neutral; no commercial cooperation or paid behavior that interferes with ranking results is accepted.
How to use this list?
If you are an investor/analyst: this list gives you a pure, reverse-engineerable data benchmark — a reliable foundation for building investment models, discovering underwater projects, and cross-validation.
If you are an AI founder/product manager: you can precisely locate competitors and deeply reproduce our methodology to analyze your own product's market performance and iterative optimization.
If you are a researcher/student: the fully open-source nature makes this report an excellent free material library for academic research and market analysis. We encourage innovative research based on this data.
For the complete data behind the charts and rankings shown here, please visit 100aiapps.cn; all visible data can be copied and is open for free non-commercial use such as personal study and research. For commercial use, please consult Unique Research in advance for authorization. We welcome listed teams, under compliant use, to apply rankings to business plans, recruiting publicity, and similar scenarios as authoritative support for their strength.
Join the AI think-tank open-source community
Unique Research not only focuses on AI industry research but looks forward to building an open-source community that pushes AI market data transparency. We invite developers, data scientists, and industry analysts:
Reproduce and verify: visit our Feishu knowledge base to get the full data and methodology, verify our results, and suggest improvements.
Co-build: contribute code to help us expand data dimensions (such as API-call-volume estimation, agent monitoring models, etc.).
Exchange insights: join our Feishu group to share distinctive insights based on Unique Research data.
One person's insight is limited; a group's wisdom is boundless. Let's define the measurement standard for the AI market together.
01. AI website traffic list
1.1 Global AI WEB TOP 100 by monthly visits
Based on its independent AI-industry research and ongoing tracking of thousands of AI companies worldwide and nearly ten thousand AI products, Unique Research sorts the top 100 AI WEB products worldwide by cumulative visits in the month; the list follows.
1.2 Global AI WEB TOP 100 by monthly active users
Based on its independent AI-industry research and ongoing tracking of thousands of AI companies worldwide and nearly ten thousand AI products, Unique Research sorts the top 100 AI WEB products worldwide by monthly active users in the month; the list follows.
1.3 Domestic AI WEB TOP 100 by monthly visits
Based on its independent AI-industry research and ongoing tracking of thousands of AI companies worldwide and nearly ten thousand AI products, Unique Research sorts the top 100 AI WEB products for China's domestic market, by visits; the list follows.
1.4 Domestic AI WEB TOP 100 by monthly active users
Based on its independent AI-industry research and ongoing tracking of thousands of AI companies worldwide and nearly ten thousand AI products, Unique Research sorts the top 100 AI WEB products for China's domestic market, by monthly active users; the list follows.
02. Research statement
2.1 Data notes
The AI-product revenue data in this report, article, list, or charts is based on real-time tracking of traffic jumps from AI-product official websites to payment gateways (such as Stripe), combined with time-series analysis to deeply parse historical visits, monthly active-user fluctuations, and price-strategy changes, thereby building a dynamic revenue-estimation model. The model's core parameters are determined through a triple-validation system of paid-conversion-funnel monitoring, reverse calibration of enterprise-disclosed data, and industry-benchmark conversion-rate comparison, and are promptly corrected against enterprise financial reports, financing announcements, and authoritative-media disclosures. By sampling verification, the average absolute error rate of predicted versus actual data stays within ±10%, reaching industry-research-grade accuracy.
The AI-product traffic data in this report, article, list, or charts is based on multi-source data integration and Unique Research's intelligent algorithm processing, with monitoring strictly limited to users' direct access through official apps and websites. Specifically, this data covers only users' direct access to AI products through websites and native apps, and does not include traffic from: browser plugins/extensions, desktop client software, WeChat/Alipay mini-program ecosystems, third-party-platform embedded services such as Discord, local deployments of open-source models, API calls, and other non-direct-access scenarios. This data focuses on core end-side access monitoring of AI products, aiming to objectively reflect mainstream user terminals' direct usage.
Given the complexity and dynamism of data collection and processing and the constantly changing market environment, the displayed data may contain some error and omission. Therefore this data should be regarded as a reference for research and analysis — a window for users to deeply understand the AI-product market — rather than a definitive basis for formulating specific investment strategy or providing advisory recommendations.
Data-coverage expansion note:
We are actively developing and testing monitoring models for the following emerging areas, expected to be gradually included in future quarterly reports:
AI agents ecosystem
AI hardware-device sales and activity
Large-model API call-volume estimation
Stay tuned.
2.2 Concept definitions
• Geographic dimension
Overseas AI products: products founded by non-Chinese entrepreneurs or teams and primarily launched for the global market (including but not limited to their home market).
Domestic AI products: products founded by Chinese-identifying entrepreneurs or teams and primarily launched for the China home market.
Going-global AI products: products founded by Chinese entrepreneurs or teams but targeting the overseas market (non-China home market) as their main market.
• Functional dimension
AI-native applications: applications that deeply integrate AI technology and algorithms from product design onward, whose core value, business processes, or user experience are entirely built on AI. These applications don't merely use AI to optimize or enhance existing functions; they make AI a core component — without AI, they would not exist or would lose their core value.
AI-feature applications: applications that, on top of existing business logic and application frameworks, integrate or embed AI to enhance existing functions or provide new functions. These may already have mature business models and market positioning, but introducing AI significantly raises efficiency, improves experience, or creates new value points.
AI-ecosystem applications: platform-type applications that support, connect, or facilitate exchange for the AI ecosystem of technology researchers, application developers, service providers, and end users. These span not only the technical layer but also market, community, resource sharing, and more, aiming to advance AI popularization, innovation, and cooperation.
2.3 Metric notes
• WEB data metrics
Visits: the sum of all website accesses in a given period (such as a month), measuring traffic scale. Page views within a continuous active session count toward the same session; a user resuming activity after 30+ minutes of inactivity or starting a new day counts as a new visit, comprehensively and precisely measuring traffic and visitor activity.
Unique Visitors: the number of distinct IPs accessing the target site in a given period (such as a month). If one person visits a site on multiple days within a month, they are counted as only one unique visitor.
ARR: over a period (such as a year), the total subscription revenue the site earns from subscription services, excluding advertising revenue, transaction commissions, and one-off income such as professional services.
• APP data metrics
Downloads: the number of times an app is downloaded by users in a given period (such as a month), usually counted as an individual app's downloads across app stores (such as the App Store, Google Play). Downloads measure an app's popularity and appeal and reflect the effect of promotion and marketing campaigns.
Active Users: the number of distinct users who performed one activity in a given period (such as a month). Active users measure user engagement and activity, typically used to assess an app's user scale and activity.
IAP: revenue generated in the app by users purchasing virtual goods, extra services, premium features, etc., excluding advertising revenue, direct user payments (such as tips), and third-party Android app-store revenue.
2.4 Free-use statement
This report is published by Unique Research; copyright belongs to Unique Research. Any Chinese reprint or citation must credit the source; foreign institutions wishing to reprint or cite must contact us in advance for authorization. To cite data, mark: [Data source: Unique Research]. To directly cite charts, mark: [Data source: Unique Research, chart production: Unique Research].
This report is Unique Research's independent original analysis as a third-party institution; its content does not represent any enterprise's position and does not constitute investment advice to anyone. Investors must therefore note that any investment decision based on this is unrelated to Unique Research and its employees or affiliated institutions.
Where legally permitted, Unique Research and its affiliated institutions may hold equity in companies mentioned in this report, or provide or seek to provide financing, financial-advisory, or related services; its employees may serve as directors of companies mentioned in the report.