我在24小时内使用ClaudeCode构建并发布了Dognames.vip。
大家好，
我是一个产品经理。作为一名不懂编码的产品经理，自己从零开始构建一个应用曾经是一个遥不可及的梦想。但通过与ClaudeCode的合作，我在一天之内成功构建并上线了Dognames。
（简单说明一下：我的母语是中文，所以这篇文章是借助AI翻译的。如果有任何不自然的表达，请多多包涵！）
以下是我的工作流程概述：
1. 发现机会：
我从SEO研究开始。使用Semrush，我发现了一个黄金关键词机会：“女孩/男孩狗名字”。这个关键词的搜索量很高，但难度（KD）较低。
2. 从想法到概念（以AI作为我的用户）：
我将这个初步想法分享给ClaudeCode。为了更深入地挖掘用户需求，我让ClaudeCode扮演潜在用户进行角色扮演。这次对话改变了游戏规则，将想法从一个简单的信息聚合网站转变为一个互动的、由AI驱动的测验，以生成名字建议。
3. 结构化产品：
接下来，我们讨论了整体产品架构，包括我计划后续构建的SEO内容系统。根据我们的讨论，ClaudeCode生成了一份详细的产品需求文档（PRD.md）。
4. 构建核心功能：
在计划确定后，我让ClaudeCode开始构建主要用户流程——测验。我们首先专注于让这个核心功能实现端到端的工作。
5. 精炼设计：
我很快意识到初始设计并不太吸引人。因此，我在Dribbble上找到了很好的设计灵感，并将其提供给ClaudeCode。它首先分析了这些参考资料，以识别设计风格、色彩搭配和关键原则。然后，我让它将这个分析整理成正式的设计风格指南，我手动进行了审核和批准。
6. 设计首页：
为了确保首页的有效性，我向ClaudeCode提供了其他优秀着陆页的设计原则。这使它清楚地了解我们需要展示哪些信息。
7. 细节迭代：
在所有参考资料的基础上，ClaudeCode完成了首页的设计。然而，一些细节仍然缺失。我继续与它合作进行润色：自定义光标、移动端响应式设计，以及寻找小狗图片（我指示它使用Unsplash或在线寻找免版税的图片并下载）。
8. 通过SEO和国际化进行扩展：
核心产品上线后，我们开始构建SEO系统并处理多语言支持（i18n）。这成为整个项目中耗时最长的部分。
9. 最终产品：
最后，你们看到的就是今天的产品。从想法到上线，整个过程不到24小时。
我分享这个案例研究是为了表明，任何愿意尝试的人都有机会构建他们所设想的产品。
我希望你们也能受到启发，去构建自己的产品。即使很少有人知道它，创造属于自己的东西总会带来一种特别的成就感，不是吗？
查看原文
Hi everyone,<p>I'm a product manager. And as a non-coding PM, the idea of building an app from scratch by myself used to be a distant dream. But by collaborating with ClaudeCode, I was able to build and launch Dognames in a single day.<p>(A quick note: My native language is Chinese, so this article was translated with the help of AI. Please excuse any awkward phrasing!)<p>Here's a breakdown of my workflow:<p>1. Discovering the Opportunity:
I started with SEO research. Using Semrush, I found a golden keyword opportunity: "girl/boy dog names." It had high search volume but low keyword difficulty (KD).<p>2. From Idea to Concept (with AI as my user):
I shared this initial idea with ClaudeCode. To dig deeper into the user needs, I had ClaudeCode role-play as a potential user. This conversation was a game-changer, pivoting the idea from a simple information aggregation site into an interactive, AI-powered quiz to generate name suggestions.<p>3. Structuring the Product:
Next, we discussed the overall product architecture, including the SEO content system I planned to build out later. Based on our discussion, ClaudeCode generated a detailed Product Requirements Document (PRD.md).<p>4. Building the Core Functionality:
With the plan in place, I had ClaudeCode start by building out the main user flow—the quiz. We focused on getting this core feature working end-to-end first.<p>5. Refining the Design:
I quickly realized the initial design wasn't very appealing. So, I found some great design inspiration on Dribbble and fed it to ClaudeCode. It first analyzed the references to identify the design style, color palette, and key principles. Then, I had it compile this analysis into a formal Design Style Guide, which I manually reviewed and approved.<p>6. Designing the Homepage:
To ensure the homepage was effective, I provided ClaudeCode with design principles from other excellent landing pages. This gave it a clear understanding of what information we needed to showcase.<p>7. Iterating on the Details:
With all the reference material, ClaudeCode completed the homepage. However, some details were missing. I continued to work with it to add polish: a custom cursor, mobile responsiveness, and sourcing puppy images (I instructed it to use Unsplash or find royalty-free images online and download them).<p>8. Scaling with SEO & Internationalization:
With the core product live, we moved on to building out the SEO system and handling multi-language support (i18n). This turned out to be the most time-consuming part of the entire project.<p>9. The Final Product:
And finally, you have the product you see today. The entire process, from idea to launch, took less than 24 hours.<p>I'm sharing this case study to show that anyone who is willing to try has the opportunity to build the product they envision.<p>I hope you're inspired to build your own product, too. Even if very few people know about it, there's a special kind of accomplishment in creating something of your own, isn't there?