我的2025：构建、暂停与寻找人们所需的产品
在2025年，我参与了多个SaaS产品的开发。大多数产品都没有成功，只有一个存活下来，并达到了600美元的月经常性收入（MRR）。
以下是我所构建的产品、失败的原因以及我所学到的内容的简要总结。
---
*1. SurelyForm*
我运营这个产品两年，注册用户超过2万，主要来自我的开源社区。我几乎没有进行任何营销。
最终只转化为2个付费客户。这些用户并不是我的真正目标受众。我失去了耐心，决定关闭它。
---
*2. Tour123*
这个产品试图解决一个真实的问题：随着产品的迭代，产品演示视频的维护成本变得昂贵。
我们的想法是从端到端的测试用例中自动生成演示视频。技术上是可行的，但没有公司愿意采用。工程师不想编写测试，产品经理说用户会跳过文档。
它解决了一个工程问题，而不是组织问题。
---
*3. Morphon*
Morphon起初是一个为客户定制的低代码平台，后来我加入了人工智能。在此期间，“氛围编码”产品迅速崛起，迅速占领市场。代币成本高，差异化不明确。我暂停了这个项目。
---
*4. Mentorbook*
Mentorbook将文件、YouTube或任何学习材料转化为针对个人目标的结构化课程。
到2025年底，它达到了600美元的月经常性收入。用户愿意付费是因为它确实帮助他们学习，而不是因为折扣或个人信任。
---
*2025年的关键教训：*
* 失败是筛选器
* 熟悉的用户 ≠ 目标用户
* 小收入与真实依赖 > 虚假的吸引力
* 知道何时停止与知道如何执行同样重要
---
我在我的博客上写了一篇更长的反思：[Mentorbook 2025反思](https://www.mentorbook.ai/blog/posts/my-2025)
查看原文
In 2025, I worked on multiple SaaS products. Most of them didn’t work. One survived and reached $600 MRR.<p>Here’s a short summary of what I built, what failed, and what I learned.<p>---<p>*1. SurelyForm*<p>I ran this product for two years. It had 20k+ registered users, mostly from my open-source communities. I did almost no marketing.<p>It converted into 2 paying customers. These users were not the real target audience. I lost patience and shut it down.<p>---<p>*2. Tour123*<p>This product tried to solve a real problem: product demo videos become expensive to maintain as products iterate.<p>The idea was to generate demo videos automatically from E2E test cases. It worked technically, but no company adopted it. Engineers didn’t want to write tests, PMs said users skip docs.<p>It solved an engineering problem, not an organizational one.<p>---<p>*3. Morphon*<p>Morphon started from a client custom low-code platform, then I added AI. During this period, “vibe coding” products exploded, quickly dominating the market. Token costs were high, differentiation unclear. I paused the project.<p>---<p>*4. Mentorbook*<p>Mentorbook turns files, YouTube, or any learning material into structured courses tailored to personal goals.<p>By the end of 2025, it reached $600 MRR. Users paid because it genuinely helped them learn, not because of discounts or personal trust.<p>---<p>*Key lessons from 2025:*<p>* Failures are filters
* Familiar users ≠ target users
* Small revenue with real dependency > fake traction
* Knowing when to stop is as important as knowing how to execute<p>---<p>I wrote a longer reflection on my blog: [Mentorbook 2025 Reflection](https://www.mentorbook.ai/blog/posts/my-2025)