reddit · SideProject · 09-15 19:50 45 分 ← 返回列表

I built a journal that reads your entries back and tells you what keeps repeating

Solo iOS dev. Spent a year on this after noticing I'd kept a journal for ages and never once learned anything from it. How it works: you write or dictate a few lines about your day, tap whether it was a good day by your own definition (you set that up front — the app never decides for you). After ~2 weeks it surfaces patterns across the notes: what drains you, what restores you, what shows up on your calmer days. Then it suggests one small thing to try, not ten. What I got wrong early:

五维评分

付费意愿

需求真实度

个人可落地

竞争空间

新鲜度

AI 分析

痛点：长期写日记却从未从中获得任何洞察——记录了很多，但没人帮你把反复出现的模式总结出来。

目标用户：有写日记/自我反思习惯的 iOS 用户，尤其是关注情绪管理、心理健康和个人成长的 20-35 岁人群

证据：Spent a year on this after noticing I'd kept a journal for ages and never once learned anything from it. / I built a streak counter and people hated it. Ripped it out. Now missed days just aren't counted — no guilt screen. / 7-day trial.

变现：订阅制（7天免费试用 + 月/年订阅），可加一次性买断的高级版；后续做 AI 周报/月报作为付费墙核心功能

MVP：2周MVP：不做完整日记App，只做一个「导入已有日记 → LLM按固定模板抽取3-5条可引用证据的模式 → 输出一张周报卡片」的Web或iOS小组件，先验证用户是否愿意为「洞察」而不是「记录」付费

竞品：Day One、Apple Journal、Rosebud、Mindsera、Reflectly、Stoic、Journey

摘要：一位独立 iOS 开发者花一年做了一个 AI 日记 App：用户写或口述几行日记并自评好坏，约两周后 App 从笔记中提取反复出现的模式（什么消耗你、什么让你恢复），并只建议一件小事。开发者还发现 streak 计数器被用户讨厌于是移除，采用 7 天试用模式，但帖子中没有任何真实收入或付费用户数据，需求也未得到其他人附和。