先决定做什么，再动手Decide what to build before you build it
你告诉我们想做什么。我们从真实用户的抱怨里核实需求有多大、在涨还是在缩、现有方案差在哪，给出“做”或“不做”，每个结论都附证据。Tell us what you're thinking of building. We check it against what real users complain about: how big the need is, whether it's growing, where the current options fall short. You get a clear build-or-don't verdict, with the evidence behind it.
这类工具有 5 万用户，一年里掉了 38%The category has 50k users and lost 38% in a year
最大的替代品还在更新，评分 4.6The biggest alternative is maintained and rated 4.6
不做。需求在缩，而且已经有人做得够好。Don't build. The need is shrinking and it's already served well.
示意：每个结论都由事实推出，事实都有出处。Illustration: every verdict follows from facts, and every fact has a source.
常见的做法是找一份“机会清单”照着做。我们核实过，大多数清单经不起检查。The usual move is to grab a list of "opportunities" and build from it. We checked. Most of those lists don't survive a closer look.
打分最高的 100 个“作者停更、仍有人用”的 Chrome 扩展，逐个联网核实后，值得重做的一个都没有。We checked the 100 top-scoring abandoned Chrome extensions one by one. Not one was worth rebuilding.
其中 90 个早就有更好、还在更新的替代品；66 个背后的需求在萎缩或已经消失。90 already had a better, maintained alternative. For 66, the underlying need was shrinking or gone.
在 30,996 个用户过千的在架扩展里，10,665 个主要是把你引到别处（注册、购物、自家产品）的入口。Of 30,996 listed extensions with 1,000+ users, 10,665 mainly exist to send you somewhere else: a signup, a shop, the maker's own app.
三层：先落到需求，再看有没有机会，最后过门槛、给出可以检验的结论。Three layers: pin the idea to a need, look for a real opening, then run it through the gates and give a verdict you can check.
把想法对应到一个真实需求，同一需求下的所有产品合起来看：多少人在用、在涨还是在缩、谁还在维护。只盯一个停更产品会看偏，上面“100 个里 0 个值得做”就是这么来的。Map the idea to a real need and look at every product serving it together: how many people use them, whether that's growing, who's still maintaining. Looking at one abandoned product in isolation is how you get the "0 out of 100" above.
有需求不等于有机会。我们从七个角度找新产品进得去的理由，首批 8 个扩展就分别来自其中不同的几种：A need isn't an opening. We look for a reason a new product can get in from seven angles. Our first 8 extensions each came from one of them:
| 上升需求Rising need | 需求一年内在涨，后来者还能起量The need is growing and newcomers still gain users |
|---|---|
| 断供Left behind | 大产品下架，原来的用户没人接住A big product disappeared and its users weren't picked up |
| 头部不满Unhappy with the leader | 头部产品的差评集中在能解决的问题上The leader's bad reviews cluster on fixable problems |
| 收费方式Better pricing | 已经有人付费，换一种收费方式能抢到用户People already pay; a different model (one-time, bring your own key) can win them |
| 产出即收入Pays for itself | 工具直接帮用户赚钱，用户不多但愿意付The tool directly earns its users money: fewer users, willing to pay |
| 平台空白Platform gap | 用户多、增长快的工作平台，工具却很少A busy, growing work platform with few tools for it |
| 捆绑Bundle | 几个互补的小需求合成一个更好用的产品Several small, related needs combined into one better product |
每个机会都要过四道门槛：政策和伦理风险（一票否决）、平台会不会自己做进去、做出来能不能长期维护、能不能拿到用户。结论是“做 / 不做”加前提条件，以及上线 4 周后用来判断成败的标准。Every opening goes through four gates: policy and ethics (an automatic no), whether the platform will build it in, whether it can be maintained, and whether you can reach users. The verdict is build or don't, the conditions, and the test you'll use 4 weeks after launch.
别人卖判断，我们公开自己判断得准不准：每种机会信号的命中率，上架结果出来后都会公开。Others sell judgment. We publish how often ours is right: the hit rate of each kind of opening, once launch results are in.
3 个工作日内交付，每一条都能追溯到证据。Delivered within 3 working days. Every point traces back to evidence.
首批 8 个 Chrome 扩展按这套方法选出、由 AI 开发。每个上架当天公开成功标准，4 周后公开对账，没达到的也照写。Our first 8 Chrome extensions were picked with this method and built with AI. Each one gets its success criteria published on launch day and a public result 4 weeks later, misses included.
| 批次Batch | 状态Status | 公开内容What's public |
|---|---|---|
| 01–03 | 已提交商店审核（2026-09-26）In store review (2026-09-26) | 上架当天公开名称和 4 周成功标准Names and 4-week targets on launch day |
| 04–08 | 已开发完成，等待发布名额Built, waiting for publishing slots | 同上Same |
现在覆盖浏览器扩展（Chrome、Edge、Firefox）和安卓、iOS 应用；其他类型的想法也可以提交，我们会告诉你能研判到什么程度。We currently cover browser extensions (Chrome, Edge, Firefox) and Android and iOS apps. Ideas of other kinds are welcome too; we'll tell you how far we can take the verdict.
一个方向，一份研判文档。试点期免费，按申请先后排队，开始研判后 3 个工作日内交付。One direction, one verdict document. Free during the pilot, queued in order of application, delivered within 3 working days of starting.
每周几个正在增长、头部用户不满意的需求，附切入点和证据。Each week, a few growing needs where users are unhappy with the leaders, with the opening and the evidence.
让 Claude、Cursor 等在动手之前，先查一下这个需求值不值得做。Let Claude, Cursor and similar tools check whether a need is worth building before they start.
只登记：正式上线时第一时间通知你。申请试点：按申请先后排队，每天研判 1 个，开始做你的研判时会发邮件告诉你。Join the list and we'll tell you the moment we launch. Apply for a pilot and you join a queue in order of application. We take one a day and email you when we start on yours.
来自公开的应用商店数据和用户评论。我们交付的是加工后的判断：规模区间、趋势、切入点和结论，不提供原始数据。Public app store data and user reviews. We deliver our analysis (size ranges, trends, openings and verdicts), not raw data.
会。所以每个结论都附证据，由你自己判断；我们也把自己判断错的例子公开。Yes. That's why every verdict comes with its evidence so you can judge for yourself, and why we publish the cases we got wrong.
不用。你只需要说想解决什么问题、给谁用，不用写具体方案；研判核实的是需求，用的都是公开数据。你提交的内容只用于给你出研判文档，不公开；交付后可以选择删除。No. Just tell us the problem and who it’s for, not your solution. We check the demand using public data. What you send is used only for your verdict, never published, and can be deleted once it’s delivered.
按登记的人想做什么来定。我们正在研究鸿蒙应用、Mac 和 Windows 桌面应用、Office 插件和 Shopify 应用的数据来源，确定能做好之前不会承诺。It depends on what people sign up to build. We're looking into data sources for HarmonyOS apps, Mac and Windows apps, Office add-ins and Shopify apps, and won't promise any of them until we can do them well.
不完全一样。苹果不公开下载量，iOS 应用的规模只能按评分数估算；文档里会标明每个结论的把握度。Not quite. Apple doesn't publish download counts, so iOS market size is estimated from ratings. Each verdict states how confident we are.
正式版会收费，价格还没定。首批登记用户享受早鸟价。The full version will be paid; pricing isn't set yet. Early sign-ups get early-bird pricing.