我的 2025：构建、暂停，以及寻找人们需要的产品
1 分•作者: tangjinzhou•6 个月前
2025 年，我同时在多个 SaaS 产品上工作。其中大部分都失败了。只有一个存活下来，并达到了 600 美元的月经常性收入（MRR）。
以下是我构建了什么、失败了什么以及我学到了什么的简短总结。
---
*1. SurelyForm*
我运营这个产品两年了。它有 2 万多注册用户，主要来自我的开源社区。我几乎没有做任何市场营销。
它转化了 2 个付费客户。这些用户并非真正的目标受众。我失去了耐心，并关闭了它。
---
*2. Tour123*
这个产品试图解决一个实际问题：随着产品迭代，产品演示视频的维护成本变得很高。
这个想法是从端到端（E2E）测试用例中自动生成演示视频。它在技术上是可行的，但没有公司采用它。工程师们不想编写测试，产品经理（PM）说用户会跳过文档。
它解决了工程问题，而不是组织问题。
---
*3. Morphon*
Morphon 起源于一个客户定制的低代码平台，后来我加入了 AI。在此期间，“氛围编码”产品迅速爆发，迅速占据了市场主导地位。代币成本很高，差异化不明显。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根据个人目标量身定制的结构化课程。
到 2025 年底，它达到了 600 美元的月经常性收入（MRR）。用户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 年的关键经验教训：*
* 失败是过滤器
* 熟悉的用户 ≠ 目标用户
* 拥有真实依赖性的小额收入 > 虚假的用户增长
* 知道何时停止与知道如何执行一样重要
---
我在我的博客上写了一篇更长的反思文章：[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)