Skip to content
Back to search

qiaomu-app-review-insights

把 App Store 评价变成产品研究证据,发现痛点、机会和版本风险 | Turn App Store reviews into product research evidence: pain points, opportunities, and version risks.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/joeseesun--qiaomu-app-review-insights/badge.svg)](/listings/joeseesun--qiaomu-app-review-insights)

HTML

<a href="/listings/joeseesun--qiaomu-app-review-insights"><img src="/api/listings/joeseesun--qiaomu-app-review-insights/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 293 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # 乔木 App 洞察
  2. > 把 App Store 用户评价变成产品经理能直接使用的洞察页:痛点、机会、版本风险、证据和可分享页面。
  3. > Turn App Store reviews into product research evidence: pain points, opportunities, version risks, and voice-of-customer signals.
  4. [![Live Demo](https://img.shields.io/badge/Live%20Demo-appreview.qiaomu.ai-0f766e?style=for-the-badge)](https://appreview.qiaomu.ai)
  5. [![Agent Skill](https://img.shields.io/badge/Agent%20Skill-qiaomu--app--review-111827?style=for-the-badge)](#agent-skill)
  6. [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
  7. [![GitHub stars](https://img.shields.io/github/stars/joeseesun/qiaomu-app-review-insights?style=social)](https://github.com/joeseesun/qiaomu-app-review-insights)
  8. ![乔木 App 洞察产品截图](docs/assets/product-screenshot.png)
  9. **中文** | [English](#english)
  10. ## 为什么值得用
  11. App Store 评论里藏着真实需求、付费阻力、版本事故和竞品机会,但它们通常散在几百条高噪声评论里。乔木 App 洞察会自动抓取评论,用分析模型提炼结构化洞察,并生成可缓存、可分享、可复盘的 App 评价洞察页面。
  12. 它适合:
  13. - 产品经理:把真实用户抱怨转成需求池和版本风险清单
  14. - 独立开发者:从竞品差评里发现小产品切入口
  15. - 增长 / 内容团队:把用户评价转成有证据的内容素材
  16. - 投研 / 行研:快速观察热门 App 的用户口碑变化
  17. - Agent 用户:用自然语言触发 App 评价分析工作流
  18. ## 核心能力
  19. | 能力 | 说明 |
  20. | --- | --- |
  21. | App 搜索 | 支持 App Store 链接、App ID、应用名称和国家区 |
  22. | 评论抓取 | Apple RSS + App Store 页面补样本,保留来源边界 |
  23. | LLM 分析 | 默认可接入 `deepseek-v4-flash`,提炼摘要、痛点、机会和行动建议 |
  24. | 证据优先 | 每个结论尽量保留代表性评论,避免空泛总结 |
  25. | 静态洞察页 | 每个 App 生成稳定 URL,记录更新时间,方便分享和复盘 |
  26. | 榜单入口 | 首页展示多国家、多分类 Top Free / Top Paid App |
  27. | 可视化诊断 | 版本口碑、差评主题、时间线和样本构成辅助判断 |
  28. | 公开保护 | 3 天内缓存复用,公开生成限流,避免 token 被打穿 |
  29. ## 样例输出
  30. 一次生成会得到:
  31. - 摘要:产品口碑和关键结论
  32. - 核心痛点:用户反复抱怨的问题,附证据句
  33. - 产品机会:可以转成路线图或独立产品想法的机会
  34. - 正向信号:用户愿意给高分的原因
  35. - 用户分层:不同用户群体的关注差异
  36. - 版本风险:更新、性能、限制、付费策略带来的风险
  37. - 行动建议:产品经理可以继续验证或排期的动作
  38. - 可视化图表:版本趋势、痛点热力、评论情绪时间线
  39. ![乔木 App 洞察图表截图](docs/assets/diagnostics-screenshot.png)
  40. 线上示例:
  41. - [ChatGPT 评价分析](https://appreview.qiaomu.ai/apps/us/6448311069/chatgpt)
  42. - [乔木 App 洞察首页](https://appreview.qiaomu.ai)
  43. ## 快速开始
  44. ```bash
  45. git clone https://github.com/joeseesun/qiaomu-app-review-insights.git
  46. cd qiaomu-app-review-insights
  47. npm install
  48. cp .env.example .env.local
  49. npm run dev
  50. ```
  51. 打开 [http://localhost:3000](http://localhost:3000)。
  52. 最小环境变量:
  53. ```env
  54. QIAOMU_LLM_API_KEY=your_api_key
  55. QIAOMU_LLM_BASE_URL=https://api.deepseek.com/v1
  56. QIAOMU_LLM_MODEL=deepseek-v4-flash
  57. NEXT_PUBLIC_SITE_URL=http://localhost:3000
  58. STORAGE_TYPE=local
  59. ```
  60. 不要把 `.env.local`、`.env.development` 或真实 API Key 提交到仓库。
  61. 公开部署建议打开生成保护:
  62. ```env
  63. APP_REVIEW_PUBLIC_DAILY_NEW_APP_LIMIT=5
  64. APP_REVIEW_PUBLIC_CACHE_FRESH_DAYS=3
  65. APP_REVIEW_GENERATION_LIMIT_DIR=/app/src/data/app-cache/.generation-guard
  66. APP_REVIEW_HISTORY_DIR=/app/src/data/app-cache/.review-history
  67. ```
  68. ## 预生成热门 App 页面
  69. ```bash
  70. node scripts/precache-top-apps.mjs \
  71. --base-url http://localhost:3000 \
  72. --countries us,cn \
  73. --charts free,paid \
  74. --limit 5 \
  75. --max-reviews 160
  76. ```
  77. 常用参数:
  78. - `--force`:覆盖已有缓存并重新抓取
  79. - `--no-analyze`:只抓评论统计,不调用 LLM
  80. - `--limit`:每个榜单生成多少个 App 页面
  81. - `--max-reviews`:每个 App 最多抓取多少条评论
  82. ## Docker 部署
  83. ```bash
  84. cp .env.example .env.production
  85. npm run build
  86. docker compose --env-file .env.production up -d --build
  87. curl http://127.0.0.1:3095/api/health
  88. ```
  89. 生产环境推荐:
  90. ```env
  91. NEXT_PUBLIC_SITE_URL=https://your-domain.example
  92. APP_REVIEW_CACHE_DIR=/app/src/data/app-cache
  93. ```
  94. 默认 Docker 端口是 `3095`。挂到 Nginx、Caddy 或 Cloudflare Tunnel 时,把外部域名代理到该端口即可。
  95. ## 核心 API
  96. 生成或读取缓存页:
  97. ```bash
  98. curl -X POST http://localhost:3000/api/research \
  99. -H 'Content-Type: application/json' \
  100. -d '{"query":"ChatGPT","country":"us","maxReviews":160}'
  101. ```
  102. 更新洞察:
  103. ```bash
  104. curl -X POST http://localhost:3000/api/research/regenerate \
  105. -H 'Content-Type: application/json' \
  106. -d '{"appId":"6448311069","country":"us","maxReviews":160}'
  107. ```
  108. 健康检查:
  109. ```bash
  110. curl http://localhost:3000/api/health
  111. ```
  112. ## Agent Skill
  113. 如果你用 Codex、OpenCode、Cursor、Cline、Warp 等支持 Agent Skills 的工具,可以安装配套 skill:
  114. ```bash
  115. npx skills add joeseesun/qiaomu-app-review-skill
  116. ```
  117. 你可以这样说:
  118. - `分析 ChatGPT 的 App Store 用户评价,重点看版本风险和产品机会`
  119. - `帮我找一个同类 App 的差评痛点,看看有没有独立开发机会`
  120. - `把这个 App 的评论整理成产品洞察页结构`
  121. Skill 会引导 Agent 使用公开站点、API、缓存页面和产品洞察分析流程。
  122. ## 技术栈
  123. - Next.js 15 + React 19 + TypeScript
  124. - Tailwind CSS
  125. - ECharts / Recharts
  126. - Apple iTunes Lookup API / App Store RSS Reviews
  127. - OpenAI-compatible SDK,默认接入 DeepSeek v4 flash
  128. - 本地 JSON 缓存,可扩展到 KV、Supabase 或对象存储
  129. - Docker Compose / standalone Next.js 部署
  130. ## 限制与边界
  131. - App Store RSS 评论接口有分页和地区差异,结果取决于国家区、抓取时间和 Apple 返回内容。
  132. - AI 洞察不是事实裁决,页面会保留原始评论证据,建议结合证据判断。
  133. - 公开生成会限流并复用缓存,避免无成本批量消耗 LLM token。
  134. - Serverless 平台需要额外配置持久化存储,否则缓存和历史评论不会稳定保留。
  135. ## Troubleshooting
  136. | 问题 | 解决方法 |
  137. | --- | --- |
  138. | `AI 服务密钥未配置或不可用` | 检查 `.env.local` 的 `QIAOMU_LLM_API_KEY` 或 `DEEPSEEK_API_KEY`,然后重启 dev server。 |
  139. | 搜索到的 App 不对 | 带上国家区和 App Store 链接,例如 `https://apps.apple.com/us/app/chatgpt/id6448311069`。 |
  140. | 详情页没有 AI 摘要 | 点击更新洞察,或确认服务能访问 LLM API。也可以先用 `--no-analyze` 只生成评论缓存。 |
  141. | Docker 启动后无缓存 | 确认 `APP_REVIEW_CACHE_DIR` 指向容器内可写路径,并挂载 volume。 |
  142. | 线上页面链接不对 | 设置 `NEXT_PUBLIC_SITE_URL=https://你的域名` 后更新缓存页。 |
  143. ## 关于向阳乔木
  144. 这个项目来自向阳乔木的 AI 工作流实践:我更关心工具如何进入真实产品、内容和增长工作,而不只是停留在模型参数或产品发布本身。
  145. - 个人站:[qiaomu.ai](https://qiaomu.ai)
  146. - 博客:[blog.qiaomu.ai](https://blog.qiaomu.ai)
  147. - 乔木推荐:[tuijian.qiaomu.ai](https://tuijian.qiaomu.ai)
  148. - X:[@vista8](https://x.com/vista8)
  149. - GitHub:[@joeseesun](https://github.com/joeseesun/)
  150. <a name="english"></a>
  151. ---
  152. # Qiaomu App Review Insights
  153. > Turn App Store reviews into product research evidence: pain points, opportunities, version risks, and voice-of-customer signals.
  154. Qiaomu App Review Insights helps product managers, indie hackers, researchers, and content teams turn noisy App Store comments into structured insight pages: summaries, pain points, opportunities, positive signals, user segments, version risks, action plans, charts, and original review evidence.
  155. ## Why It Exists
  156. App Store reviews contain product truth: why users churn, why they pay, which updates broke trust, and where competitors leave openings. Reading them manually is slow. This project fetches review samples, runs LLM-assisted analysis, and generates durable pages that are easy to share, revisit, and cite.
  157. ## Highlights
  158. - Search by App Store URL, App ID, or app name
  159. - Fetch reviews from Apple RSS and App Store page samples
  160. - Analyze reviews with an OpenAI-compatible LLM, defaulting to `deepseek-v4-flash`
  161. - Preserve review evidence and source boundaries
  162. - Generate cached, stable insight pages per app
  163. - Show Top Free / Top Paid apps by country and category
  164. - Visualize version sentiment, review sources, pain themes, and time ranges
  165. - Protect public deployments with cache reuse, rate limits, and a generation queue
  166. ## Local Development
  167. ```bash
  168. git clone https://github.com/joeseesun/qiaomu-app-review-insights.git
  169. cd qiaomu-app-review-insights
  170. npm install
  171. cp .env.example .env.local
  172. npm run dev
  173. ```
  174. Open [http://localhost:3000](http://localhost:3000).
  175. Minimum environment variables:
  176. ```env
  177. QIAOMU_LLM_API_KEY=your_api_key
  178. QIAOMU_LLM_BASE_URL=https://api.deepseek.com/v1
  179. QIAOMU_LLM_MODEL=deepseek-v4-flash
  180. NEXT_PUBLIC_SITE_URL=http://localhost:3000
  181. STORAGE_TYPE=local
  182. ```
  183. ## API
  184. Generate or read a cached page:
  185. ```bash
  186. curl -X POST http://localhost:3000/api/research \
  187. -H 'Content-Type: application/json' \
  188. -d '{"query":"ChatGPT","country":"us","maxReviews":160}'
  189. ```
  190. Refresh insights:
  191. ```bash
  192. curl -X POST http://localhost:3000/api/research/regenerate \
  193. -H 'Content-Type: application/json' \
  194. -d '{"appId":"6448311069","country":"us","maxReviews":160}'
  195. ```
  196. Health check:
  197. ```bash
  198. curl http://localhost:3000/api/health
  199. ```
  200. ## Agent Skill
  201. Install the companion skill:
  202. ```bash
  203. npx skills add joeseesun/qiaomu-app-review-skill
  204. ```
  205. Example prompts:
  206. - `Analyze ChatGPT App Store reviews and focus on version risks.`
  207. - `Find product opportunities from a competitor's negative reviews.`
  208. - `Turn this App Store page into a product insight report.`
  209. ## License
  210. MIT

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.