July 22, 11:13

Tencent releases Miora—the first full-featured design agent platform after closed beta

The Tencent WorkBuddy team has introduced Miora, its first specialized design agent platform. Following the completion of the closed beta testing period, the platform is now available to all users at miora.design.

Functionality and Use Cases

Miora is a full-featured platform covering areas such as branding, user interface (UI) development, e-commerce, and interactive content creation. It supports multimodal functionality and an extensive library of skills. This platform expands Tencent's portfolio of artificial intelligence products beyond coding and work agents into the realm of creative design.

The platform covers five main design domains: branding, film and creative, e-commerce advertising, interactive games, and web applications. Each domain has its own set of skill templates and workflows. The interface follows the WorkBuddy design pattern and features a high degree of polish. A sidebar on the left contains modules for creation, inspiration, skills, and memory. The central chat area allows users to input design requests in natural language.

Models and Operation Configuration

Next to the Miora mascot is a scene switcher that enables instant switching between domains and dynamically updates the library of shortcut templates. At the bottom is a configuration panel where two levels of the agent's base model are available: Standard (for simple tasks), Pro (for most use cases), and Max (for complex brand identity projects). There is also a separate multimodal generation panel, allowing users to select individual models for images, video, UI, and 3D, including GPT-image-2, Seedance 2.0, Kling, Vidu Q3, and others.

Skills System and Strategic Significance

Miora's skills system transforms general design workflows into executable templates. For example, in branding mode, users can activate skills to create a complete visual brand case, generate brand characters, or design logos. The skills architecture is designed to be extensible: each template launches a vertically configured task agent that manages the process from requirements analysis to result generation. Users can choose 'Ask mode,' where the agent confirms the design direction before execution, or automatic execution to speed up iterations.

Market Positioning for Agents

Tencent's entry into the design agent market is a strategic move amid growing competition among code and video agents. The design agent sphere is defined as one of four key agent categories, alongside work, code, and video agents. Tencent holds significant advantages due to the WorkBuddy infrastructure and user base, Tencent's extensive ecosystem of models, including the HY foundational models, and partnerships with leading multimodal generation providers, as well as deep experience in product design, which is reflected in Miora's polished user experience.

Current Limitations and Prospects

Like other agent platforms, Miora faces the 'last mile' problem: AI-generated designs still require manual vectorization and professional refinement in tools like Adobe Illustrator before final use. Currently, the platform works with raster image formats but does not support vector formats, including SVG, PSD, and AI. Nevertheless, Miora marks an important step toward democratizing professional design capabilities, lowering the barrier for startups and small teams to create quality identity that previously required hiring agencies or senior designers. Tencent's foray into the design agent segment indicates a consolidation of the platform level of the AI agent ecosystem around several major operating systems for agents, upon which specialized domain agents are then built.

pandaily.comJuly 15, 05:04

Tencent transforms into an AI agent factory, expanding its ecosystem with products like WorkBuddy

By 2026, Tencent is showing signs of transforming into a powerful agent production hub. The company launched WorkBuddy, which attracted 8.85 million monthly visitors, and introduced a range of new products into its ecosystem, including Zaohua Gongfang, Ardot, Miora, WorkRally, DataBuddy, and LearnBuddy.

Strategic Shift in AI Development

Tencent, which was perceived as lagging in AI product implementation over the past two years, has adjusted its strategy. Instead of focusing on creating a single super-app with AI input, the company has begun the process of 'agentification' across its entire product ecosystem. This has led to the formation of an extensive portfolio of agentic products covering areas from desktop productivity and design to data processing and education, making Tencent one of the most active developers of AI agents in China.

Product Functionality and Integration

WorkBuddy integrates virtually all capabilities of the Tencent ecosystem as skills, including WeChat, Tencent Docs, Tencent Meeting, and enterprise services, weaving over two decades of product networking experience into a grid accessible by the agent. Currently, this product leads in the category of desktop agents in China based on daily active users and total user count.

New Directions in the Ecosystem

In addition to WorkBuddy, Tencent has released a wave of new agentic products. Zaohua Gongfang is a platform for collaborative character and story creation. Ardot functions as an AI equivalent of Figma for intelligent design, while Miora is a creative design agent. WorkRally is an AI-based comic creation platform. TDream focuses on creating interactive films and game resources using AI. DataBuddy provides AI-based data engineering and analysis services, and LearnBuddy offers expert-guided self-learning.

Difference in Approach from Competitors

Tencent's method differs from competitors' approaches because the agents are embedded within existing product ecosystems rather than requiring users to switch to a new application. For example, the AI assistant Dayuan was integrated into Enterprise WeChat, Tencent Docs evolved towards real-time collaborative text writing with AI, and Tencent Drive now supports petabyte-level asset management using AI. Tencent Lexiang acts as the DNA of corporate knowledge. These improvements transform users' familiar workflows into an agent-oriented experience without requiring behavioral changes.

Structural Prerequisites for the Strategy

The main structural factor underlying Tencent's strategy is that the company already owns some of the densest digital entry points in China: WeChat, Enterprise WeChat, QQ, Tencent Docs, Tencent Meeting, and numerous gaming platforms. Instead of creating a new super-app to compete with ByteDance Doubao or Baidu ERNIE Bot, Tencent can allow agents to evolve directly within these existing scenarios.

The Main Challenge Facing the Company

The biggest problem is not creating individual agent products, but ensuring their functionality across Tencent's organizational boundaries. The question arises: will an agent be able to cross departmental, product line, and isolated data storage boundaries? Will it be able to call capabilities from various Tencent systems on behalf of a single user? If successful, the Agent will become a new operational layer across the entire Tencent ecosystem. Otherwise, each Agent will remain just another tool in an increasingly overloaded product portfolio. According to Pony Ma, the company's path resembled a 'leaking ship' last year, then managed to stay afloat, and now is undergoing an urgent acceleration. The 'agent factory' strategy reflects Tencent's attempt to make speed its competitive advantage.