Tencent Launches Standalone Yuanbao Search Webpage, Moving Into Baidu's Search Heartland
On November 11, Tencent launched a standalone Yuanbao Search webpage. Users can search webpages, WeChat Official Accounts, WeChat Channels, and academic literature without downloading the app, and have AI organize the answers. The article notes that Yuanbao Search had previously opened to enterprises and developers, with an index scale in the hundreds of billions and hundreds of millions of daily calls, supporting more than 50 businesses including WorkBuddy and ima. The author argues that Tencent is entering search through its own content ecosystem, while Baidu is facing pressure from declining online marketing revenue after AI transformed search.
Why read this
AI score 55
Starting from Tencent Yuanbao Search's standalone webpage, this compares the AI search transformation pressure facing Baidu and Google, and presents the competitive logic of search entry points and content ecosystems.
Topics
Body · AI translation
Article | Beyond the Page, by Huahua
Everyone is watching the Agent wars, but Tencent has quietly stepped into Baidu's backyard.
On November 11, Tencent launched Yuanbao Search, a standalone web page. Users can search web pages, official accounts, Video Accounts, and academic literature without downloading an app, and can also have AI organize the answers.
This will directly hit Baidu's core territory.
First, consider a set of easily overlooked facts.
As of December 2025, China's search engine user base stood at 782 million, accounting for 69.5% of all internet users. Although this figure has declined from 878 million at the end of 2024, search remains an internet market with hundreds of millions of users.
But the question is, why is Yuanbao doing this?
When Yuanbao was first created, it existed as an important bargaining chip for Tencent in competing for the AI entry point. Since this year, the advancement of WorkBuddy has extended the focus of competition from AI assistants to office work, knowledge management, and Agent execution.
But Agents have a practical threshold: users must first have a task before they think of using one.
More often, people simply want to know one thing, clarify one question, or find one piece of material. Checking the weather, learning about a company, verifying a piece of news—these needs arise every day, and users already know where to go to address them.
For Yuanbao, chat and Agents have expanded the boundaries of its capabilities, while search provides a usage scenario that is easier to trigger.
This is also what makes the standalone web entry point noteworthy.
If search were merely a feature within the Yuanbao app, it would first serve users who have already entered Yuanbao; by setting up a separate web page, users can use it directly from a browser, and the scope of competition expands accordingly.
In 2025, Baidu accelerated the AI transformation of search. By July of that year, the proportion of Baidu mobile search results pages containing AI-generated content had reached 64%, and in October this proportion rose further to about 70%.
The changes go beyond the results page.
Baidu upgraded its search box into an intelligent box that supports long text and multimodal input; search results expanded from traditional link lists to structured content such as text, images, videos, and files, and it also integrated MCP services so that users can invoke external tools.
Looking at the data, Baidu's user base remains enormous. In the second quarter of 2026, Baidu App's monthly active users reached 644 million. Web page indexing, search algorithms, and user habits are all accumulations that latecomers find difficult to replicate in the short term.
The commercial chain of traditional search is relatively clear. Users enter keywords, the platform displays web pages and ads, users click through to websites, and advertisers pay for traffic. AI integrates information into the results page, users may no longer open the original web pages, and the advertising model that relied on clicks is therefore under pressure.
This is not simply a product upgrade issue; it is a business that Baidu must personally change—the very business it relies on to make money.
In the second quarter of 2026, Baidu's online marketing revenue was 13.1 billion yuan, down 19% year on year. Although this change is affected by multiple factors such as macroeconomic advertising demand and industry competition, and cannot be simply attributed to AI search, the pressure on the traditional advertising business makes the commercial pressure of search transformation more evident.
This is not a problem for Baidu alone; Google on the other side of the ocean is also undergoing similar changes.
III. Google wants to preserve the web ecosystem
How did Google do it?
After launching AI Overviews, Google introduced AI Mode, allowing users to ask follow-up questions around a single issue. In June 2026, Google disclosed that AI Overviews had more than 2.5 billion monthly active users and AI Mode had more than 1 billion monthly active users.
In the third quarter of 2026, Google's search advertising revenue grew 12% year on year to $63.4 billion, below market expectations and markedly slower than the nearly 20% growth earlier this year.
Google Chief Business Officer Philipp Schindler acknowledged on the earnings call that AI Mode is affecting the traditional search business and that advertising monetization capacity is being suppressed.
There is also a very important difference between Google and Baidu.
What Baidu first needs to solve is how to transform search advertising revenue, whereas what Google faces is not only its own advertising business but also the sustainability of the entire web ecosystem.
Google has long connected websites, news organizations, professional content producers, and advertisers. Search engines obtain information from web pages while sending users back to websites through links. This mechanism allows content producers to gain traffic and gives search engines a continuously updated source of information.
It is precisely against this backdrop that Google made a noteworthy adjustment.
On October 15, Google announced adjustments to how AI Mode and AI Overviews are presented, increasing the visibility of web links in answers so that important sources are easier to click.
This is not simply a return to traditional search. The value of links has already changed. Previously, they were mainly responsible for bringing users to relevant web pages, but today they also serve to mark sources, verify answers, and maintain content supply.
By raising the visibility of links again, Google is in effect rearranging the relationship among models, users, and content producers.
And this happens to point to another card in Tencent's hand: content.
IV. Tencent's hidden card is the content ecosystem
In 2006, Tencent launched Soso; in 2013, the Soso business was merged into Sogou, and thereafter search capabilities gradually settled into products such as WeChat Search. In the era of traditional web search, Baidu had already accumulated an enormous web index, algorithms, and user habits, and it was difficult for Tencent to move the market with just a search box.
This time, re-entering the market, Tencent has changed its approach: instead of attacking head-on from a full-web index, it starts from its own thickest content pool.
In October 2026, Tencent Cloud had already opened Yuanbao Search to enterprises and developers. This search system has an index scale of hundreds of billions and daily average call volume reaching hundreds of millions of times, supporting more than 50 businesses including WorkBuddy, ima, Tencent Financial Technology, and Tencent Video, and has also connected 157 authoritative sources across 13 vertical domains such as finance, healthcare, law, and travel.
This means that before Yuanbao Search launched for ordinary users, it was already running in multiple business scenarios inside Tencent, accumulating technical and service capabilities.
But the content ecosystem card is attractive, yet it is not easy to actually play.
The volume of content on WeChat Official Accounts and Channels is enormous, but the quality is uneven. Research from professional institutions, experience from industry practitioners, personal opinions, and marketing content may all appear around the same keyword.
A rigorous industry analysis and a rewritten marketing account piece may appear in adjacent search results. The more content there is, the heavier the responsibility of filtering becomes.
Another unavoidable issue is commercialization.
Li Ruizhang, product lead for the Tencent Yuanbao Search web version, said in a media interview that Yuanbao Search does not reject advertising, but users should not have to look for answers inside ads.
This statement actually touches the most sensitive boundary of AI search: when ads may also be integrated by the model into answers, the platform must keep commercial information clearly distinguished from organic information. Users can accept ads in search results, but it is hard for them to accept being unable to tell where an answer actually comes from.
The content ecosystem gives it the chips to enter the game, but it cannot directly win users for it.
The content ecosystem is Tencent's hidden card for re-entering the market, while the standalone web page determines where it starts competing for users.
But the value of search is not only in obtaining information. Without stable information retrieval capabilities, it is difficult for Agents to continuously complete work that requires real-time information.
This is also why Tencent Cloud had previously opened Yuanbao Search for products such as WorkBuddy and ima to use. Search does not have to exist only as a feature of Yuanbao; it can become an underlying capability called by multiple Agents together.
Looking again at Tencent's product layout, several lines begin to connect: Yuanbao serves individual users, WorkBuddy takes on office tasks, ima handles knowledge management, the WeChat ecosystem provides content, and search provides information retrieval capabilities for these products.
This also explains why Tencent wants to make search into a standalone web page. It hopes both to serve users' most everyday information needs and to extend search capabilities into more complex Agent tasks.
Looking back at the development of the internet, competition over entry points has always been changing.
In the PC era, search engines brought users to web pages; in the mobile era, super apps aggregated content and services into their own ecosystems. In the AI era, search has begun to filter information and organize answers for users, and Agents then turn information into action.
Each time the entry point changes, the relationship between platforms and users is rewritten.
The more complete the answer, the more easily users stop asking further questions. But often, the truly important information is hidden precisely where the answer does not present it.