Hirelytics is a small full-stack React app (Vite + React Router) that provides AI-powered resume analysis and ATS scoring. Users upload a PDF resume, the app converts it to an image, sends the file to an AI feedback service (via the Puter.js integration), stores results in a KV store, and displays a detailed review.

• Upload PDF resumes (drag & drop)
• Convert the first page of a PDF to an image preview
• AI-generated structured feedback (ATS score, content, style, structure, skills)
• Persistent storage via the integrated puter client (KV, fs)
• View and share analyzed resumes with a unique review page
• Lightweight server build for SSR using React Router
• Dockerfile for containerized deployment

• Frontend: React 19 + TypeScript
• Routing: React Router (v7) with SSR support
• Bundler: Vite
• Styling: Tailwind CSS
• State: Zustand (used for puter store)
• PDF handling: pdfjs-dist to render PDF pages to canvas
• AI & storage: puter.js (injected via script tag) — auth, fs, kv, and AI endpoints

• Node.js 20+ and npm
• A puter.js runtime available in the browser (the app loads https://js.puter.com/v2/ in the HTML layout)

Install dependencies:

npm ci

Run the dev server (Vite + React Router dev server):

npm run dev

Build for production:

npm run build

Start the production server (serves build/):

npm run start

Notes:

• The app uses the puter browser SDK to access auth, file upload, AI, and KV features. Ensure the https://js.puter.com/v2/ script can initialize in your environment.

Build and run the included Dockerfile:

docker build -t hirelytics .
docker run -p 3000:3000 hirelytics

The Dockerfile creates a multi-stage build and runs the built server from build/server.

1. User uploads a PDF on the Upload page (/upload). See app/routes/upload.tsx.
2. The file is uploaded via puter.fs.upload, and the app converts the first page to an image using app/lib/pdf2img.ts.
3. The app sends the uploaded PDF path and prepared instructions (job title & description) to puter.ai.chat via puter's feedback helper. The expected AI response is a JSON object matching the Feedback interface (see constants/index.ts).
4. The resulting feedback is stored in the KV store (puter.kv.set) under resume:<uuid>.
5. The user is redirected to /resume/:id to view the review, rendered with components in app/components.

• Layout & app entry: app/root.tsx
• Routes config: app/routes.ts
• Upload flow: app/routes/upload.tsx
• Resume review: app/routes/resume.tsx
• Puter client wrapper (Zustand store): app/lib/puter.ts
• PDF conversion helper: app/lib/pdf2img.ts
• AI instructions & format: constants/index.ts
• Dockerfile: Dockerfile

• The app expects puter.js to be available in the browser. The layout includes this script tag: <script src="https://js.puter.com/v2/"></script> (see app/root.tsx).
• No additional server-side environment variables are required for the built-in example; production deployments may integrate a proper puter backend or runtime.

• Linting/type generation: npm run typecheck runs react-router typegen and tsc.
• The Vite config uses @react-router/dev and Tailwind Vite plugin: vite.config.ts.

• Open an issue or submit a PR. Keep changes focused and add tests where appropriate.

• Add CI (GitHub Actions) to run npm ci and npm run typecheck.
• Add e2e tests for the upload and review flow.
• Secure production puter credentials / runtime if deploying publicly.