Back to search
qiaomu-app-review-insights
SkillGitHub
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
Markdown
[](/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)
- # 乔木 App 洞察
- > 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。
- > Turn App Store reviews into product research evidence: pain points, opportunities, version risks, and voice-of-customer signals.
- [](https://appreview.qiaomu.ai)
- [](#agent-skill)
- [](LICENSE)
- [](https://github.com/joeseesun/qiaomu-app-review-insights)
- 
- **中文** | [English](#english)
- ## 为什么值得用
- 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。
- 它适合:
- - 产品经理:把真实用户抱怨转成需求池和版本风险清单
- - 独立开发者:从竞品差评里发现小产品切入口
- - 增长 / 内容团队:把用户评价转成有证据的内容素材
- - 投研 / 行研:快速观察热门 App 的用户口碑变化
- - Agent 用户:用自然语言触发 App 评价分析工作流
- ## 核心能力
- | 能力 | 说明 |
- | --- | --- |
- | App 搜索 | 支持 App Store 链接、App ID、应用名称和国家区 |
- | 评论抓取 | Apple RSS + App Store 页面补样本,保留来源边界 |
- | LLM 分析 | 默认可接入 `deepseek-v4-flash`,提炼摘要、痛点、机会和行动建议 |
- | 证据优先 | 每个结论尽量保留代表性评论,避免空泛总结 |
- | 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 |
- | 榜单入口 | 首页展示多国家、多分类 Top Free / Top Paid App |
- | 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 |
- | 公开保护 | 3 天内缓存复用,公开生成限流,避免 token 被打穿 |
- ## 样例输出
- 一次生成会得到:
- - 摘要:产品口碑和关键结论
- - 核心痛点:用户反复抱怨的问题,附证据句
- - 产品机会:可以转成路线图或独立产品想法的机会
- - 正向信号:用户愿意给高分的原因
- - 用户分层:不同用户群体的关注差异
- - 版本风险:更新、性能、限制、付费策略带来的风险
- - 行动建议:产品经理可以继续验证或排期的动作
- - 可视化图表:版本趋势、痛点热力、评论情绪时间线
- 
- 线上示例:
- - [ChatGPT 评价分析](https://appreview.qiaomu.ai/apps/us/6448311069/chatgpt)
- - [乔木 App 洞察首页](https://appreview.qiaomu.ai)
- ## 快速开始
- ```bash
- git clone https://github.com/joeseesun/qiaomu-app-review-insights.git
- cd qiaomu-app-review-insights
- npm install
- cp .env.example .env.local
- npm run dev
- ```
- 打开 [http://localhost:3000](http://localhost:3000)。
- 最小环境变量:
- ```env
- QIAOMU_LLM_API_KEY=your_api_key
- QIAOMU_LLM_BASE_URL=https://api.deepseek.com/v1
- QIAOMU_LLM_MODEL=deepseek-v4-flash
- NEXT_PUBLIC_SITE_URL=http://localhost:3000
- STORAGE_TYPE=local
- ```
- 不要把 `.env.local`、`.env.development` 或真实 API Key 提交到仓库。
- 公开部署建议打开生成保护:
- ```env
- APP_REVIEW_PUBLIC_DAILY_NEW_APP_LIMIT=5
- APP_REVIEW_PUBLIC_CACHE_FRESH_DAYS=3
- APP_REVIEW_GENERATION_LIMIT_DIR=/app/src/data/app-cache/.generation-guard
- APP_REVIEW_HISTORY_DIR=/app/src/data/app-cache/.review-history
- ```
- ## 预生成热门 App 页面
- ```bash
- node scripts/precache-top-apps.mjs \
- --base-url http://localhost:3000 \
- --countries us,cn \
- --charts free,paid \
- --limit 5 \
- --max-reviews 160
- ```
- 常用参数:
- - `--force`:覆盖已有缓存并重新抓取
- - `--no-analyze`:只抓评论统计,不调用 LLM
- - `--limit`:每个榜单生成多少个 App 页面
- - `--max-reviews`:每个 App 最多抓取多少条评论
- ## Docker 部署
- ```bash
- cp .env.example .env.production
- npm run build
- docker compose --env-file .env.production up -d --build
- curl http://127.0.0.1:3095/api/health
- ```
- 生产环境推荐:
- ```env
- NEXT_PUBLIC_SITE_URL=https://your-domain.example
- APP_REVIEW_CACHE_DIR=/app/src/data/app-cache
- ```
- 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。
- ## 核心 API
- 生成或读取缓存页:
- ```bash
- curl -X POST http://localhost:3000/api/research \
- -H 'Content-Type: application/json' \
- -d '{"query":"ChatGPT","country":"us","maxReviews":160}'
- ```
- 更新洞察:
- ```bash
- curl -X POST http://localhost:3000/api/research/regenerate \
- -H 'Content-Type: application/json' \
- -d '{"appId":"6448311069","country":"us","maxReviews":160}'
- ```
- 健康检查:
- ```bash
- curl http://localhost:3000/api/health
- ```
- ## Agent Skill
- 如果你用 Codex、OpenCode、Cursor、Cline、Warp 等支持 Agent Skills 的工具,可以安装配套 skill:
- ```bash
- npx skills add joeseesun/qiaomu-app-review-skill
- ```
- 你可以这样说:
- - `分析 ChatGPT 的 App Store 用户评价,重点看版本风险和产品机会`
- - `帮我找一个同类 App 的差评痛点,看看有没有独立开发机会`
- - `把这个 App 的评论整理成产品洞察页结构`
- 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。
- ## 技术栈
- - Next.js 15 + React 19 + TypeScript
- - Tailwind CSS
- - ECharts / Recharts
- - Apple iTunes Lookup API / App Store RSS Reviews
- - OpenAI-compatible SDK,默认接入 DeepSeek v4 flash
- - 本地 JSON 缓存,可扩展到 KV、Supabase 或对象存储
- - Docker Compose / standalone Next.js 部署
- ## 限制与边界
- - 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。
- - 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。
- - 公开生成会限流并复用缓存,避免无成本批量消耗 LLM token。
- - 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。
- ## Troubleshooting
- | 问题 | 解决方法 |
- | --- | --- |
- | `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。 |
- | 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`。 |
- | 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。 |
- | 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。 |
- | 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。 |
- ## 关于向阳乔木
- 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。
- - 个人站:[qiaomu.ai](https://qiaomu.ai)
- - 博客:[blog.qiaomu.ai](https://blog.qiaomu.ai)
- - 乔木推荐:[tuijian.qiaomu.ai](https://tuijian.qiaomu.ai)
- - X:[@vista8](https://x.com/vista8)
- - GitHub:[@joeseesun](https://github.com/joeseesun/)
- <a name="english"></a>
- ---
- # Qiaomu App Review Insights
- > Turn App Store reviews into product research evidence: pain points, opportunities, version risks, and voice-of-customer signals.
- 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.
- ## Why It Exists
- 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.
- ## Highlights
- - Search by App Store URL, App ID, or app name
- - Fetch reviews from Apple RSS and App Store page samples
- - Analyze reviews with an OpenAI-compatible LLM, defaulting to `deepseek-v4-flash`
- - Preserve review evidence and source boundaries
- - Generate cached, stable insight pages per app
- - Show Top Free / Top Paid apps by country and category
- - Visualize version sentiment, review sources, pain themes, and time ranges
- - Protect public deployments with cache reuse, rate limits, and a generation queue
- ## Local Development
- ```bash
- git clone https://github.com/joeseesun/qiaomu-app-review-insights.git
- cd qiaomu-app-review-insights
- npm install
- cp .env.example .env.local
- npm run dev
- ```
- Open [http://localhost:3000](http://localhost:3000).
- Minimum environment variables:
- ```env
- QIAOMU_LLM_API_KEY=your_api_key
- QIAOMU_LLM_BASE_URL=https://api.deepseek.com/v1
- QIAOMU_LLM_MODEL=deepseek-v4-flash
- NEXT_PUBLIC_SITE_URL=http://localhost:3000
- STORAGE_TYPE=local
- ```
- ## API
- Generate or read a cached page:
- ```bash
- curl -X POST http://localhost:3000/api/research \
- -H 'Content-Type: application/json' \
- -d '{"query":"ChatGPT","country":"us","maxReviews":160}'
- ```
- Refresh insights:
- ```bash
- curl -X POST http://localhost:3000/api/research/regenerate \
- -H 'Content-Type: application/json' \
- -d '{"appId":"6448311069","country":"us","maxReviews":160}'
- ```
- Health check:
- ```bash
- curl http://localhost:3000/api/health
- ```
- ## Agent Skill
- Install the companion skill:
- ```bash
- npx skills add joeseesun/qiaomu-app-review-skill
- ```
- Example prompts:
- - `Analyze ChatGPT App Store reviews and focus on version risks.`
- - `Find product opportunities from a competitor's negative reviews.`
- - `Turn this App Store page into a product insight report.`
- ## License
- MIT
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.