Resume Analyzer (Full Stack)

A small full-stack app that analyzes a resume (pasted text or uploaded file) and returns scores + structured feedback as JSON.

How it works

1. Frontend (Vite + React) collects either:
   • Pasted resume text, or
   • A file upload (PDF / JPG / PNG / WEBP)
2. Frontend sends a request to the backend:
   • Text mode: JSON body { resumeText, targetRole }
   • File mode: multipart/form-data with file and targetRole
3. Backend (Express):
   • Extracts text from PDF (PDF.js) or runs OCR on images (Tesseract)
   • Classifies whether the document looks like a resume (LLM)
   • If it is a resume, asks an LLM to produce JSON-only feedback:
     • relevanceScore, atsScore, toneScore, readabilityScore
     • missingKeywords, overallFeedback, suggestions, strengths, weaknesses
4. Frontend renders the returned JSON into a UI (score rings, sections, keyword tags).

Project structure

• frontend/: React UI (Vite dev server)
• backend/: Express API that extracts/ocr’s text and calls the model

Quickstart (Windows / PowerShell)

Backend

cd backend
npm install
npm run start

Backend runs on http://localhost:3000 by default.

Frontend

Open a second terminal:

cd frontend
npm install
npm run dev

Frontend runs on http://localhost:5173 by default.

Configuration (environment variables)

This repo uses .env files in backend/ and frontend/.

• Backend: see backend/README.md for required variables.
• Frontend: see frontend/README.md for variables and backend URL wiring.

Important: keep secrets (API keys) out of git history. Don’t paste real keys into documentation.

API overview

• POST /analyze
  • Text mode: send JSON { "resumeText": "...", "targetRole": "..." }
  • File mode: send multipart/form-data with file=<pdf|image> and optional targetRole

If the uploaded file is not recognized as a resume, the backend responds 400 with a message and a short reason.