Baidu Miaoda: 500K Commercial Apps, 5 Billion Yuan Value

The concept of "Vibe Coding," recognized by the British "Collins English Dictionary" as a word of the year, has seen rapid global adoption in 2025. This trend, which allows users to create applications by expressing intentions in natural language rather than writing code, has spurred significant investment and product development.

Major capital injections include Cursor's valuation nearing $9.9 billion and Google's acquisition of key Windsurf members to launch Antigravity. Product offerings have expanded with Claude Code, v0, and Lovable internationally, alongside Trae (ByteDance), Qcoder (Alibaba), and Comate (Baidu) in China. New large language models like Gemini 3 and GPT 5.2 have also been used to "Vibe" websites shortly after their release.

While social media has showcased numerous Vibe Coding demonstrations, many remain in the prototype stage, often lacking the robust features required for commercial applications.

From Prototypes to Production-Grade Applications

Traditional software engineering has progressed through increasing levels of abstraction, from machine language to high-level programming languages. Low-code platforms simplified development but still demanded a "programmer's mindset." Vibe Coding aims to remove this limitation by focusing on user intent, making natural language the new compiler.

However, for an application to be considered "production-grade," it typically requires persistent storage, state management, service integration, and high availability. Many current Vibe Coding products excel at generating front-end interfaces but often lack the back-end logic, data persistence, and multi-user interaction capabilities essential for commercial viability.

Ant Group's Lingguang, launched this year, emphasizes "Flash Apps" for rapid development. Baidu's Miaoda platform, introduced nearly a year earlier, has shifted its focus from simple "toy making" to building production-grade applications. An IDC report identifies Miaoda as a significant player in the AI no-code sector, highlighting its "platform capability" and "application quality," particularly in "service integration" and "application complexity support."

Miaoda's international version, MeDo, recently partnered with Supabase, a global open-source back-end service provider, becoming one of its initial partners for a new white-label solution. This collaboration aims to enhance Miaoda's full-stack capabilities.

Commercial Impact and Ecosystem

At its 2025 Creator Conference, Miaoda reported that its platform has generated over 500,000 commercial applications within eight months of launch. These applications, with a daily new application growth exceeding 150%, cover more than 200 scenarios, including education, business, and enterprise services. The platform claims to have created over 5 billion yuan in economic and efficiency value.

These applications are described as commercially viable products, enabling creators to develop and monetize solutions without extensive coding knowledge.

Miaoda's production-grade capabilities are attributed to Baidu's self-developed front-end and back-end infrastructure and specialized model capabilities. The platform uses multiple AI agents to simulate a development team, starting with generating a product requirements document from user prompts. This process addresses potential oversights or inaccuracies in initial user ideas, ensuring a comprehensive plan before development.

The platform integrates code, plugins, and back-end services, allowing users to call plugins and configure back-end services through dialogue. Its self-developed models and databases ensure performance, while agents automate database generation, queries, and back-end services, simplifying development.

Miaoda also features a plugin ecosystem, including pre-installed Baidu services like Search and Maps, an AI tool matrix for content generation, and support for third-party and custom plugins. Users can integrate external capabilities using simple "@" commands, reducing the need for manual API integration.

Real-World Applications and Monetization

Examples of applications built on Miaoda include an oil and gas well production optimization design system, which has been deployed in major oilfields and educational institutions, reportedly saving over 1.4 million yuan in costs. Another project, an office tool designed to display novel text while appearing to type, demonstrates the platform's flexibility for creative solutions.

Miaoda supports the distribution of applications as WeChat mini-programs or native applications and leverages Baidu's search engine for visibility. The platform has seen over 20,000 applications integrate payment capabilities, completing over 80,000 transactions.

Miaoda applications primarily fall into three categories: direct monetization tools like e-commerce mini-programs, business software for internal systems, and AI efficiency applications. These have collectively generated over 5 billion yuan in economic and efficiency value. Developers have used Miaoda to create complex systems, such as customized enterprise management solutions, with significantly reduced development time and cost compared to traditional methods.

One developer, Li Jianfei, noted that Miaoda's ability to accurately interpret requirements and generate UI, logic, and databases integrally led his team to choose it over other AI programming tools.

Miaoda's product evolution, from simple application creation to production-grade development, reflects a focus on underlying infrastructure. The platform aims to simplify technical concepts like databases, APIs, and authentication by embedding them within its agent-driven interface, allowing users to focus on business logic.

The platform's "Dream Building Plan," launched on December 16, aims to support 1 million creators in achieving commercial revenue over the next three years and cultivate 1,000 developers to generate an accumulated annual revenue increase of 50 million yuan within one year. It also offers investment opportunities for high-potential projects.