China Digital Digest Weekly: Exploring the Chinese Digital Landscape
Hi folks, we are back with our weekly edition of China’s Digital Digest, wherein we bring you weekly updates on China’s digital space. The report takes a quick glance at China’s complex and rapidly evolving social media landscape by providing updates on the latest happenings across the social media industry. Here are the major highlights of the report.
1. Alibaba to Integrate AI Across Taobao and Tmall Merchant Operations
Alibaba plans to expand the use of artificial intelligence across merchant operations on its Taobao and Tmall e-commerce platforms in 2027.
Tmall President Liu Bo said the initiative will build on existing AI tools used for product launches, advertising and customer service. A key component is Qwen Work, an enterprise AI assistant designed to support merchants operating across multiple platforms. Alibaba is also integrating AI into shopping recommendations and product discovery to help merchants reach more customers and drive sales.
2. Alibaba Launches QwenWork E-Commerce Tool for Multi-Platform Sellers
Alibaba has introduced an e-commerce edition of its AI assistant, QwenWork, to help online merchants manage operations across multiple platforms.
Currently in invite-only beta testing in China, the tool features a unified dashboard for monitoring sales, inventory and advertising performance. It can also analyse social media trends, recommend marketing strategies and generate product images and promotional copy in bulk. The platform integrates with Alibaba’s merchant tools and third-party services. However, it cannot yet independently manage stores or adjust advertising campaigns.
3. Doubao Phone Assistant Expands AI Automation to Third-Party Apps
ByteDance’s Doubao Phone Assistant is gradually expanding its screen automation capabilities to third-party apps after the public consultation period for its Screen Automation Execution Protocol (SAEP) ended. The protocol allows app developers to permit or restrict AI-driven operations within their services.
However, 151 apps, including WeChat, Taobao and Alipay, have opted out of the feature. Developers can still refuse access at any time. The move highlights growing tensions between AI agents seeking cross-app functionality and platforms prioritising security, privacy and control over their services.
4. ByteDance’s Doubao Work Adds AI Data Migration Tools
ByteDance’s Doubao Work has introduced a data import and synchronisation feature that allows users to transfer preferences, skills, memory, projects and conversation histories from ChatGPT, Claude Code and TRAE Work.
Imported conversations retain their original context, enabling users to continue existing tasks without reconfiguring tools or repeating background information. The update also introduces a “Continue in New Chat” option, message annotations and a conversation-summary sidebar. These additions aim to simplify cross-platform workflows and help users manage complex projects more efficiently.
5. Tencent’s Toast App Allows External APK and H5 Imports for Pro Users
Tencent’s AI app-building platform Toast has introduced a feature allowing Pro subscribers to import existing Android applications and HTML5 (H5) source code. Developers can upload APK packages created using other tools or import H5 projects to continue editing within Toast.
The platform also streamlines app descriptions, screenshots, compliance documentation and submission to Tencent’s MyApp store. Previously, its publishing workflow primarily supported applications created within Toast. The update extends these capabilities to existing projects, helping independent developers and startups prepare their products for distribution without rebuilding them from scratch.
6. Tencent Open-Sources Hy-Memory 2.0 to Improve AI Agent Memory
Tencent Hunyuan has open-sourced Hy-Memory 2.0, an AI memory framework designed to help agents retain information across conversations and long-running tasks.
The release includes a dedicated memory model, a framework for managing stored information and MemHub, an application that synchronises memories across multiple AI tools. The framework supports three types of memory: chat, working and coding. These capabilities help agents track user preferences, retain task progress and remember why code was changed. MemHub also supports conversation-history imports and continuous synchronisation across connected tools.
7. Tencent Launches Marvis AI Assistant on Kylin Software Store
Tencent has launched its Marvis AI assistant on the Kylin Software Store, expanding its availability across China’s domestic operating system ecosystem.
The software supports Galaxy Kylin Desktop OS V10 SP1 on x86 and ARM architectures, covering multiple domestic processors and mainstream computing platforms. Marvis operates at the system level, allowing users to adjust computer settings, install or uninstall software, perform security checks and diagnose technical issues through natural-language commands. It also offers local document search, task automation and remote PC control through a mobile app, targeting productivity needs across government agencies and state-owned enterprises.
8. Tencent Launches Free WorkBuddy AI Education Edition for Teachers
Tencent has launched a free education-focused version of its WorkBuddy AI assistant for teachers across China. Eligible educators can receive 2,000 credits monthly for 12 months, regardless of their location or teaching level.
Developed by Tencent’s Sustainable Social Value (SSV) division and the WorkBuddy team, the platform offers AI teaching assistants designed to support lesson planning, presentation creation, differentiated assignments, parent communication and student development. It also includes tools for creative teaching, scripted lessons, animation demonstrations and educational content production. Future integrations with Tencent Docs and other services aim to streamline teachers’ daily workflows.
9. Rokid Partners With Tencent to Bring Voice-Based Ride-Hailing to AI Glasses
Rokid has partnered with Tencent Mobility to enable users to book rides through its AI glasses using voice commands.
Connected through an Agent-to-Agent (A2A) protocol, Rokid’s assistant can work with Tencent’s mobility agent to suggest pickup points, check nearby vehicles, compare fares, place bookings and process password-free payments. Users can also view vehicle details, license plates and estimated arrival times through the glasses. The feature aims to simplify ride-hailing without requiring users to repeatedly check their phones. However, phones remain a fallback for handling payment issues, destination changes and cancellations.
10. WeChat for HarmonyOS Integrates Huawei’s Celia Assistant Through A2A
WeChat’s HarmonyOS app has integrated Huawei’s Celia voice assistant through an agent-to-agent (A2A) connection, allowing users to send messages and initiate voice or video calls using voice commands.
With WeChat version 8.0.22 and Celia version 11.7.8.303 or later, users can dictate messages to individual contacts and confirm actions before they are executed. The feature does not yet support group chats, group calls, stickers or WeChat’s cloned-account function. The integration highlights how A2A technology can connect system-level AI assistants with third-party apps, enabling users to complete everyday tasks through voice commands without navigating each app manually.
11. DeepSeek Opens Search and Fetch APIs to AI Agent Developers
DeepSeek has introduced Search and Fetch APIs that allow AI agents to search the web and retrieve webpage content without relying on separate search providers.
Search returns structured results for queries, while Fetch extracts webpage content in Markdown format for use by AI models. Both APIs cost RMB 5 per 1,000 successful requests, with failed requests and empty search results excluded from billing. Integration with the existing DeepSeek API also simplifies authentication for developers. The launch could reduce reliance on external search services, although it raises questions about data processing, provider dependence and search quality compared with established alternatives.
12. Xiaomi Open-Sources AI Model That Can Identify Where Sounds Originate
Xiaomi has open-sourced MiDashengLM-Spatial, an audio-language model designed to understand both sound content and its spatial location.
Released under the Apache 2.0 licence, the model combines general audio understanding with the ability to identify sound sources and reason about their positions and spatial relationships. Built on Xiaomi’s MiDashengLM-7B, it adds a dedicated spatial audio encoder while retaining capabilities such as speech recognition, audio captioning and environmental sound analysis. Xiaomi plans to bring the technology to smart-home products and vehicle cabins, potentially enabling AI assistants to respond more precisely based on what they hear and where sounds originate.
Wrapping Up
The vast and diverse nature of the Chinese Social Media space makes it incredibly challenging to keep a tab on the rapid developments taking place. However, China’s Digital Digest brings you all the latest updates from there to keep you abreast of all the evolving trends.
To delve deeper into the findings of our latest report, click here.
Comments