AI Codes a BI Dashboard and Two WeChat Mini-Programs in One Month
For solo developers, the code is now the easy part; WeChat's certification fee and ICP filing are what actually block a launch. This month of AI-assisted building shows that platform compliance, not coding, is the new cost center for mini-program development.
Over one month, AI-based code generation enabled a solo developer to ship a Java BI analytics tool called SuperBI, a WeChat mini-program for course schedules that is now live, and a second mini-program for workplace wellness still in the ICP filing queue. SuperBI connects to MySQL, SQLite, Oracle, and SQL Server, runs read-only SELECT queries, and presents data through tables, bar and line charts, and numeric cards. Its pipeline runs from login and permissions through data source connection, metadata reading, SQL query, and visualization configuration. The tool relies on Codex for code generation and is currently hosted on Gitee, with a standalone version planned because source deployment via java -jar still has a high barrier.
The course schedule mini-program was built entirely by WeChat's DevTools after entering a one-line description, running on Tencent Cloud native server and database. It has cost nothing so far, but the author hesitates over the 30-yuan WeChat certification fee, suspecting WeChat is monetizing the surge of AI-generated mini-programs. The third app, named "Workstation Idle" after many name rejections, passed WeChat review and is awaiting ICP filing, after which it should go live automatically. WeChat requires certification within a month or the mini-program loses searchability.
The whole exercise shows that coding is no longer the bottleneck for small tool projects; platform compliance and certification fees now dominate the remaining friction.
The biggest practical cost of AI-assisted development is shifting from writing code to navigating platform rules like ICP filing and WeChat certification.
The decision to keep SuperBI on Gitee rather than GitHub suggests the tool targets a domestic Chinese audience, but also reveals a cautious stance on open-sourcing AI-generated code.
A 30-yuan fee is trivial in absolute terms, but when coding is essentially free, such platform fees become the only real obstacle to launch, which may be exactly why WeChat charges them.