---
title: "Emotional Companionship Outperforms Everyone: Is This the Real Star of AI Going Overseas?"
author: "Unique Research"
sourcePublication: "Unique Research Substack"
originalPublishedAt: "2026-05-17T12:58:47+00:00"
canonical: "https://ffcap.cn/en/research/src-20260517-01html"
source: "https://uniqueresearch.substack.com/p/src-20260517-01html"
language: "en"
---

# Emotional Companionship Outperforms Everyone: Is This the Real Star of AI Going Overseas?

_Original · Unique Research · 2026-05-17_

_Editor's note: This is a monthly data report by Unique Research covering April 2026 AI-going-overseas web-side traffic. All visit counts, active-user counts, month-over-month percentages, rankings, and market-share figures are Unique Research's monitoring data or third-party data attributed in the source, not independently audited findings. Product and company names (PolyBuzz, manus, KLING AI, Meshy, Butterfly Effect, Zhipu AI, Wondershare Technology, etc.) are retained as working English renderings where official English forms remain unverified. The manus acquisition block and founder travel-restriction reports are attributed to Reuters and The Washington Post; the KLING AI US$300 million annualized revenue figure is a reported claim. The data scope is limited to direct web and native-app access; browser extensions, desktop clients, mini-programs, embedded third-party services, local open-source model deployment, and API calls are excluded. The source is dated May 17, 2026, covering April 2026 data._

Global AI Application Rankings

What Happened in AI Going Overseas in April

Unique Research · April 2026 AI-Going-Overseas Web-Side Monthly Report Interpretation

"

Every month we run web-side data on overseas AI products. Not all tracks are still growing—some products have declined to a frown-inducing degree; other tracks are quietly building energy. This content lays out the data and talks about it.

The Emotional Companionship Track Is Still That Strong

Number one in April visits was not Qwen, not manus—it was still PolyBuzz, the emotional-companionship product under Zuoyebang.

Monthly visits of 54.97 million, nearly 60% more than second-place Alibaba's Qwen (34.72 million). Among global AI-going-overseas products, only a few can compare to this number.

Zuoyebang, in everyone's印象, is an education company. But PolyBuzz is another line—a conversational AI companionship product, taking the role-play and emotional-communication route. This product's visits are now more than 7 times those of Gauth (7.6 million).

The logic behind this isn't complicated: in the emotional-companionship track, user stickiness is extremely high, daily-active duration is easy to build up, and the ceiling for paid conversion is much higher than for many tool-type products.

And this track hasn't reached the top yet. MiniMax's Talkie had 5.81 million visits in April, up 34.7% month-over-month; FlowGPT's Emochi had 11.04 million visits, up 14%. Across the entire emotional-companionship track, the top few products are growing in sync, indicating that demand is real and incremental is sustained.

One overseas user review's wording I remember clearly: PolyBuzz's multi-role chat rooms and personality-memory function "make it stand out among competitors." This isn't to say it's perfect, but it has genuinely grasped a real need of overseas users—not looking for a tool, but looking for a character.

manus: Data Didn't Collapse, but the Story Is Complicated

manus was number one in active users this month, at 9.86 million, up 9.8% month-over-month. Visits of 39 million, third in the overall rankings.

Considering what happened in April, this number is somewhat unexpected.

At the end of April, the Chinese government formally blocked Meta's proposed US$2 billion acquisition of manus. Before that, founders Xiao Hong and Ji Yichao were reportedly restricted from leaving the country, and the whole matter was heatedly discussed in overseas media. Reuters, The Washington Post, and Reuters all reported extensively. (Note: source lists Reuters twice; preserved as-is.)

But the data didn't collapse.

There are two explanatory directions: one, manus's user base is mainly overseas professional users, who care about whether the product can help them complete complex tasks and don't much care about the company's equity ownership; two, this incident objectively brought manus a round of free global exposure, and some users may have learned about the product for the first time because they saw the news.

Either way, from a business perspective, manus remains the most presentable card among Chinese going-overseas products in the global agent track. Butterfly Effect's other product, Monica, had 7.87 million visits and is also growing steadily (+1.7%). Both of this company's products are on a positive track.

Zhipu AI's Z.ai agent data is also impressive. Visits of 12.22 million, +18.6% month-over-month; active users 3.11 million, +23.3% month-over-month. Dual-line growth, and from a not-small base—quite forceful.

KLING AI "Collapsed"?

This is the most eye-popping number in this data: KLING AI, monthly visits fell from a previous base to 3.95 million, -71.7% month-over-month; active-user decline -65.6%.

Both metrics double-killed simultaneously, and both declines are in the 60–70% range—this doesn't look like natural churn, more like an active strategic adjustment.

KLING AI was originally a product with a very bright future. In January this year, reports said its annualized revenue had reached US$300 million, which is globally top-tier in the entire AI video-generation track. Foreign media wrote an article titled "Sora Closed, KLING Made 300 Million," about the logic that application layer makes more money than model layer in the AI video-generation track.

Then this month its traffic dropped 70%.

KLING AI 3.0 was released in February, with high technical reviews, supporting 4K, native lip-sync, and audio generation—word of mouth isn't bad. But as Seedance 2.0 and Happy Horse heat up, it's possible KLING AI is cooling off. But whatever the reason, a decline of this magnitude is a bit surprising, and we'll continue to watch.

Video Generation: Not Cooling Off, but Shuffling the Deck

Some people seeing KLING's decline might think the video-generation track has gone cold, but the data tells a different story.

Of the top three in the growth rankings, two are video generation:

Pexo, +1871%, 350,000 visits;

Flova, +114%, 530,000 visits;

Vozo (Shijian Ruilai), +44.6%, 540,000 visits.

Pexo's 1871% is a statistical "startup surge"—because last month's volume was too small, the multiple looks exaggerated, but the direction is real. Flova and Vozo's numbers are much more substantial; their absolute volumes are already not small, and this growth rate indicates users are genuinely using them.

Additionally, the veteran product HeyGen is also growing, with 11.25 million visits, +5.9%, and 4.36 million active users, +9.3%. In video generation, products that vertically target commercial scenarios (such as marketing videos, product videos) have much wider moats than general video generation. HeyGen's user willingness to pay and retention areobvious better than competitors taking the pan-entertainment route.

So the accurate description of video generation is not "gone cold," but "the top is reseating itself." Previously front-runners KLING and Pollo.ai (-30%) are under pressure, while new players are rising rapidly.

3D Modeling: The Most Underrated Track

We mention it every time we release a report, but this track genuinely has nobody talking about it.

Meshy, the 3D-modeling product under Taichi Graphics, had 6.09 million visits this month, with active users up 27.2% month-over-month—one of the fastest-growing among top products. Hitem3D AI, active users +18.1% month-over-month. MOOTION, visits +25.9% month-over-month.

Three products growing simultaneously, and from not-small bases.

The logic behind this is: demand for AI-generated 3D content is being pulled simultaneously by several industries—gaming, e-commerce product display, AR/VR content production, architecture and industrial design. Among these demands, some have already formed paid closed loops.

Compared with emotional companionship and video generation, 3D modeling has much less voice, but conversely, competition is also not as fierce, and products that break out first will get a more stable position.

A Few Small Tracks Quietly Rising

PPT generation: AiPPT visits 790,000, +27.5%; PPT.AI visits 750,000, +16.3%. This track's ceiling isn't high, but the scenario is extremelyrigid, and Chinese companies' product capability here is genuinely leading.

Math education AI: Mathos AI (+22.4%) and Solvely (+18.7%) are both growing, while ByteDance's Gauth dropped 11.9%. An interesting contrast—new products vertically focused on math problem-solving are grabbing traffic from older products.

Audio generation: Noiz AI visits 690,000, +35%. Fish Audio (Hanabi AI), 3.93 million visits, +18.6%. These two products one large one small, but direction is consistent—AI audio generation is finding its user base.

Object recognition: PictureThis (Ruiqi Software), visits 1.99 million, +43.5%, active users 1.31 million, +46%. This product mainly does plant recognition, with a solid overseas user base and considerable growth.

Looking at Who's Winning at the Company Level

Butterfly Effect: One of the best-performing going-overseas companies currently. manus number one in active users, Monica steady growth—two products协同, firmly established in both agent and writing-assistant tracks.

Zuoyebang: Pulled off a beautiful turnaround with PolyBuzz. Nobody expected an education company to take global first place in the emotional-companionship track, but they did it.

Wondershare Technology: The multi-product matrix playbook has worked. Filmora, OpusClip, Media.io, UniConverter—several products covering different video-generation scenarios, UniConverter +27.3% this month, overall steady with progress.

ByteDance: Severe internal differentiation. Dola AI (personal assistant) +21%, Dreamina (image generation) +6.6%, both not bad. But KLING AI's big drop is a difficult problem, and Gauth is also declining. ByteDance has laid out a wide product line going overseas, but each product's state is uneven.

Alibaba: Qwen is the strongest open-source player in chatbots globally, with 34.72 million visits, firmly second in the rankings. But Accio (AI-driven B2B procurement search) dropped 45.5% this month—although it rose sharply in March, April's decline is a bit惨 and worth watching.

How the Overseas World Views Chinese AI Going Overseas

A few data points can help build macro perception.

Chinese AI products' global market share rose from ~1% at the end of 2024 to ~15% in November 2025. In February this year, on OpenRouter, the world's largest AI model aggregation platform, Chinese models' API call volume surpassed the US for the first time—5.16 trillion tokens vs the US's 2.7 trillion in one week, a considerable gap, and 47% of that platform's users are American.

In a16z's global AI consumer app Top 100 released in March this year, manus, KLING AI, and Qwen all made the list—the most Chinese products ever in a single edition.

But there's also a sentence foreign media has repeated many times, to the effect of: Chinese AI applications' user data privacy concerns make some overseas users hesitant. This isn't groundless—PolyBuzz's overseas user review explicitly wrote "privacy protection is a mixed bag." Data trust is currently an unavoidable threshold for Chinese going-overseas products, especially for the European market with high compliance requirements.

MiniMax listed on the Hong Kong stock exchange in January this year, and its market cap is now more than 4x its IPO price; the IPO prospectus states overseas revenue accounts for over 70%—this is the most overseas-oriented Chinese AI company. Its Talkie visits rose 34.7% this month. This company is worth continued observation.

A Few Final Judgments

The biggest insight from this data is that the application layer is the real battlefield.

Competition at the model layer is between China and the US, and most people can't participate. But opportunities at the application layer are open—whoever can deeply and well serve a real user need can find their place on this ranking.

Emotional companionship, agents, 3D modeling, PPT generation, math education AI—most products in these tracks don't have the highest technical barriers, but getting users willing to pay and willing to stay requires understanding demand and patient product refinement.

Video generation will continue to shuffle; the old order is breaking, and the new top hasn't been determined.

Emotional companionship's ceiling is still far, but competition will become increasingly fierce.

KLING AI's problem deserves Kuaishou's serious attention; a 70% single-month decline is no small matter.

We'll continue following next month.

Data source: Unique Research · Statistical caliber: monthly web-side visits, April 2026

About Us

Unique Research - Authoritative Institution for Global AI Market Insights

The details of Unique Research's April 2026 AI-Going-Overseas Web Rankings are as follows:

AI-Going-Overseas Website Traffic Rankings: Top 100 rankings of visits and monthly active users for overseas AI web products.

All ranking lists simultaneously provide growth rankings,convenient investment institutions discover potential projects, and entrepreneurs and developers discover potential opportunities.

Our Influence and Recognition:

Unique Research's data and insights have gained widespread trust and citation from top domestic and international investment institutions, academic units, and media, becoming an important benchmark for their market analysis, investment decisions, and trend judgment. Our partners and citation parties include (but are not limited to):

Securities and investment banks: Sinolink Securities, CITIC Construction Investment, Huayuan Securities, Haitong Securities, etc.

Authoritative media: China Securities Journal, TMTPost, Jiemian News, etc.

We firmly believe that only data that can withstand verification by core industry participants truly has long-term value.

We believe in the power of open source and openness, and Unique Research hereby makes the following three commitments:

This AI product ranking is permanently open source and free to access, with complete methodology and raw data supporting full reproduction.

This AI product ranking's data sources are strictly limited to independent third-party monitoring platforms; enterprise self-reported data is not accepted as an evaluation basis.

This AI product ranking's ranking system maintains absolute neutrality; no commercial cooperation or paid behavior that interferes with ranking results is accepted.

How to Use This Ranking?

If you are an investor/analyst: This ranking provides you with a clean data benchmark that can withstand reverse-engineering verification, a reliable cornerstone for building investment models, discovering underwater projects, and cross-validation.

If you are an AI entrepreneur/product manager: You can not only precisely locate competitors, but also deeply reproduce our methodology for analyzing your own product's market performance and optimization iteration.

If you are a researcher/student: The fully open-source nature makes this report an excellent free material library for your academic research and market analysis. We encourage innovative research based on this data.

For the complete data behind the charts and rankings shown in this article, please visit 100aiapps.cn. All visible data can be copied and is open for free personal learning, research, and other non-commercial use; for commercial scenario use, please consult Unique Research for authorization in advance. We welcome ranked teams to apply the rankings to business plans, recruitment promotions, and other scenarios under compliant use, providing authoritative evidence of your strength.

Join the AI Think Tank Open Source Community

Unique Research not only focuses on AI industry research, but also looks forward to building an open source community that promotes AI market data transparency. We invite developers, data scientists, and industry analysts to:

Reproduce and verify: Visit our Feishu knowledge base to obtain complete data and methodology, verify our results, and propose improvements.

Co-build: Contribute code to help us expand data dimensions (such as API call volume estimation, agent monitoring models, etc.).

Overseas AI Web Rankings

1.1 Overseas AI Web TOP 100 by Monthly Visits Ranking

Based on its independent research of the AI industry and continuous tracking of thousands of AI companies and nearly 10,000 AI products globally, Unique Research sorts by the cumulative visits of that AI web product in the current month, and compiles the top 100 AI web products by Chinese or Chinese entrepreneurs targeting overseas markets, sorted by visits. The specific ranking is as follows:

1.2 Overseas AI Web TOP 100 by Monthly Active Users Ranking

Based on its independent research of the AI industry and continuous tracking of thousands of AI companies and nearly 10,000 AI products globally, Unique Research sorts by the active user count of that AI web product in the current month, and compiles the top 100 AI web products by Chinese or Chinese entrepreneurs targeting overseas markets, sorted by active users. The specific ranking is as follows:

Research Statement

2.1 Data Description

The AI product revenue data involved in this report, article, ranking, or chart is based on real-time tracking of traffic redirected from AI product official websites to payment gateways (such as Stripe), combined with time-series analysis, deeply analyzing historical visits, monthly active user fluctuations, and pricing strategy changes, to construct a dynamic revenue estimation model. The model's core parameters are determined through a triple-verification system of paid-conversion-funnel monitoring, reverse calibration of enterprise-disclosed data, and comparison of industry benchmark conversion rates, and are promptly compared and corrected with corporate financial reports, financing announcements, and authoritative media disclosures. Through sampling verification, the average absolute error rate between predicted values and actual data is stable within ±10%, with accuracy reaching industry-research standards.

The AI product traffic data involved in this report, article, ranking, or chart is based on multi-source data integration and Unique Research's intelligent algorithm processing, with monitoring scope strictly limited to users' direct access behavior through official applications and websites. Specifically, this data only covers users' direct access to AI products through websites and native applications, and does not include traffic data generated by the following access methods: browser plugins/extensions, desktop client software, WeChat/Alipay mini-program ecosystems, embedded services on third-party platforms such as Discord, local deployment of open-source models, API interface calls, and other non-direct-access scenarios. This data focuses on monitoring core end-side access behavior of AI products, aiming to objectively reflect direct usage of mainstream user terminals.

Given the complexity and dynamics of data collection and processing, as well as the constantly changing market environment, the displayed data may contain a certain degree of error and omission. Therefore, this data should be regarded as a reference basis for research and analysis, aiming to provide users with a window for deeply understanding the AI product market, rather than a definitive basis for directly formulating specific investment strategies or providing consulting advice.

Data Coverage Expansion Note:

We are actively developing and testing monitoring models for the following emerging fields, expected to be gradually included in future quarterly reports:

AI agent (Agents) ecosystem

AI hardware device sales and activity

Large-model API call scale estimation

Stay tuned.

2.2 Concept Definitions

• Geographic Dimension

Overseas AI products: refers to products founded by non-Chinese entrepreneurs or teams, and primarily targeting the global market (including but not limited to their home market).

Domestic AI products: refers to products founded by Chinese entrepreneurs or teams, primarily targeting the Chinese domestic market.

Going-overseas AI products: refers to products founded by Chinese entrepreneurs or teams, but positioning the main market target overseas (non-Chinese domestic market).

• Functional Dimension

AI-native applications: refers to applications that from the beginning of product design deeply integrate artificial intelligence technology and algorithms, whose core value, business processes, or user experience completely rely on AI technology construction. These applications not only use AI to optimize or enhance existing functions, but take AI as their core component; without AI technology, these applications would cease to exist or lose their core value.

AI-functional applications: refers to applications that on the basis of original business logic and application frameworks, enhance existing functions or provide entirely new functions by integrating or embedding AI technology. These applications may originally have had mature business models and market positioning, but through introducing AI technology, can significantly improve efficiency, enhance user experience, or create new value points.

AI-ecosystem applications: refers to platform-type applications that provide support, connection, or promotion of communication for the AI ecosystem composed of AI technology researchers, application developers, service providers, and end users. These applications are not limited to the technical level, but involve multiple dimensions such as market, community, and resource sharing, aiming to promote the popularization, innovation, and cooperation of AI technology.

2.3 Metric Definitions

• Web Data Metrics

Visits: refers to the sum of all visit counts of a website within a specific time period (such as one month), used to measure the scale of website traffic. Page views within a continuous active period are counted in the same session, while when a user reactivates after an interval of more than 30 minutes or a new day begins, it is regarded as a new visit, thereby comprehensively and accurately measuring website traffic and visitor activity.

Unique Visitors: refers to the number of independent IPs visiting the target website within a specific time period (such as one month). If someone visits a website on multiple days within a month, they are only counted as one unique visitor.

Subscription Revenue (ARR): refers to the total revenue sum obtained by a website through providing subscription services to users within a certain period (such as one year), and this revenue does not include advertising revenue, transaction commissions, and professional services with one-time characteristics.

• App Data Metrics

Downloads: refers to the number of times an App is downloaded by users within a specific time period (such as one month). Usually measured by the number of downloads of a single App on app stores (such as App Store, Google Play, etc.). Downloads are one of the important metrics for measuring an App's popularity and user appeal, and can also reflect the effectiveness of App promotion and marketing activities.

Active Users: refers to the number of independent users who have performed one activity within a specific time period (such as one month). Active users is a metric for measuring user participation and activity level in a specific time period, usually used to evaluate an App's user scale and user activity.

In-App Purchase Revenue (IAP): refers to revenue generated by users' purchase behavior in the App to obtain virtual goods, additional services, advanced features, etc., excluding advertising revenue, direct user payments (such as tipping, etc.), and revenue from third-party Android app stores.

2.4 Disclaimer

This report is published by Unique Research, and the copyright of this report belongs to Unique Research. Any Chinese reprint or citation must indicate the report source; foreign institutions wishing to reprint or cite must contact for authorization in advance. If citing data, please mark: \[Data source: Unique Research\]; if directly citing charts, please mark: \[Data source: Unique Research, Chart production: Unique Research\].

This report is an independent original analysis by Unique Research as a third-party institution. The content of the report does not represent the position of any enterprise, and does not constitute investment advice to anyone. Therefore, investors must note that any investment decisions made based on this are unrelated to Unique Research and its employees or affiliated institutions.

To the extent permitted by law, Unique Research and its affiliated institutions may hold equity in companies mentioned in the report, or provide or seek to provide related services such as fundraising or financial advisory services, and their employees may serve as directors of companies mentioned in the report.

---

Original publication: https://uniqueresearch.substack.com/p/src-20260517-01html
On-site reading page: https://ffcap.cn/en/research/src-20260517-01html
