Comic Workshop Hits 138 Users: A One-Month Solo Build and the VibeCoding Bugs That Came With It

Comic Workshop Launch Passes 100 Users: Full Retrospective of Public Account Comics from Need to Monetization

Hello everyone, I'm Xiao Feichang, a former programmer working hard to run a one-person company. Comic Workshop went live on November 4, and users have already reached 138. From October to now, the development cycle has been one month, with 10+ versions iterated. Thanks to my warm little assistant Claude. Today I'll tell you about my process from finding a need, to developing a tool, to a successful launch.

1. Comic Workshop is in public beta

This tool took over a month from project initiation in October to launch. Now it can handle the entire comic production process, including topic library construction, benchmarking material organization, one-click replication of benchmark comics, matrix operation, and other features.

Detailed feature descriptions can be found at:

Today's article mainly tells you how I discovered the need and then turned it into monetization.

2. My VibeCoding journey, from need discovery to monetization landing

2.1. Why I made the WeChat public account Comic Workshop

I've been working on this WeChat public account comic project since last September. The complete path went from Coze workflow to n8n, then to skill, and finally to Comic Workshop. Why make Comic Workshop? The core is targeting the novice market. Previously, doing WeChat public account comics required a certain level of technical foundation to a greater or lesser extent. Many novices got stuck at the configuration stage, which caused the training camps to frequently polarize: technically savvy group members had already started producing comics and monetizing, while novices were still stuck at the configuration steps.

So I made up my mind to integrate all comics into a single Comic Workshop, covering topic library construction, benchmarking material organization, one-click replication of benchmark comics, and support for matrix operation.

2.2. Pitfalls I encountered in VibeCoding

After a month of development, I also discovered some pitfalls in VibeCoding. VibeCoding used to be called AI programming. I first used it in 2024, back when it was only for writing module code. Now it can manage the entire development process.

Full management also has its troubles. VibeCoding has a fatal flaw: as the development cycle iterates, when AI fixes bugs it often causes "A is fixed, B breaks." I reviewed this with the AI, and the core reason is that it is "lazy." Tell it A has a bug, it fixes A; the AI usually does not consider the connected logic after A is modified. For example, if the user module reports an error, AI fixes it, but doesn't pay attention to whether the connected modules will have problems as a result. So when we do VibeCoding, we must follow a rule: high cohesion, low coupling.

Another point: to keep VibeCoding running smoothly, try to use a better model. I used the full-power Claude throughout.

3. Later plans

After Comic Workshop is refined, my plan is to land the mini-program project. Previously I ran a Xiaohongshu training camp, where the monetization rate was over 50%, so I extended it with a session of mini-program training camp, planning to use Xiaohongshu as the traffic entry and mini-program for monetization.

After the mini-program is completed, it's story-driven e-commerce and dynamic comics. Recently I've been trying to integrate Liaozhai-themed comic drama into the Comic Drama Workshop, expected to go live in November. Recent debugging has shown some progress, purely using skills, and still needs tuning.

That's today's article. Thanks for watching. See you next time.