AI Decision System
Replayable, offline-verifiable records for AI-assisted go/no-go decisions: experiment promotions, typed model answers and human-approved actions.
翻译准备中
尚未检测到更新。GitHub 发布版本后会显示在这里。
个人 GitHub 账号下公开仓库的所有者可以登录并验证所有权,自行管理介绍,无需管理员批准。
登录并生成验证文件,在 30 分钟内将其提交到此仓库的默认分支。其他分支或分叉仓库中的文件无效。然后在下方检查文件,即可无需管理员批准直接管理项目。请保留文件以维持管理权限。
如需登录,将前往登录页面。登录后请回到此处,再次点击所选的验证按钮。
在项目文件夹中运行 /nomorevibe verify,即可验证所有权并自行管理介绍。
一分钟完成验证VitalScan - AI-powered Web Vitals analyzer using PageSpeed Insights
翻译准备中
取自仓库 README
正在加载完整 README。 以下为已保存的摘录。
AI-powered web performance analyzer that scans any URL for Core Web Vitals and provides actionable recommendations using Claude.
• Full CWV Audit — Scans LCP, INP, CLS, FCP, TBT, Speed Index, and TTFB
• Mobile + Desktop — Side-by-side analysis for both strategies
• AI Recommendations — Claude analyzes results and gives prioritized, actionable fixes
• Visual Dashboard — Color-coded score gauges, metric cards, and opportunity list
• One-click Deploy — Ready for Vercel
Open http://localhost:3000.查看完整 README ↗AI-powered web performance analyzer that scans any URL for Core Web Vitals and provides actionable recommendations using Claude.
git clone https://github.com/YOUR_USERNAME/web-vitals-analyzer.git
cd web-vitals-analyzer
npm install
cp .env.example .env
Edit .env with your API keys:
GOOGLE_PSI_API_KEY=your_key_here
ANTHROPIC_API_KEY=your_key_here
Getting your API keys:
| Key | Where | Cost |
|---|---|---|
GOOGLE_PSI_API_KEY | Google Cloud Console → Enable "PageSpeed Insights API" → Create Credentials → API Key | Free |
ANTHROPIC_API_KEY | Anthropic Console → Create Key | Pay-per-use |
npm run dev
Open http://localhost:3000.
npx vercel
Or connect your GitHub repo in the Vercel Dashboard. Set the two environment variables in your Vercel project settings.
src/
├── app/
│ ├── page.tsx # Main page with URL input
│ ├── layout.tsx # Root layout
│ ├── globals.css # Global styles + Tailwind
│ └── api/
│ ├── analyze/ # POST: Fetches PSI data (mobile + desktop)
│ └── recommend/ # POST: Sends results to Claude for analysis
├── components/
│ ├── Dashboard.tsx # Main results dashboard with tabs
│ ├── ScoreGauge.tsx # Circular SVG score indicator
│ ├── MetricCard.tsx # Individual metric display card
│ └── AnalysisReport.tsx # Renders Claude's markdown analysis
└── lib/
├── psi.ts # Google PageSpeed Insights API client
└── claude.ts # Claude analysis prompt & client
User enters URL
→ POST /api/analyze
→ Calls PSI API (mobile + desktop in parallel)
→ Returns structured metric data
→ POST /api/recommend
→ Formats PSI data into prompt
→ Sends to Claude for expert analysis
→ Returns markdown recommendations
→ Dashboard renders scores, metrics, opportunities, and AI analysis
MIT