CV Generator (Next.js + Tailwind)

This project is a resume generator that:

• Scrapes a job description (uses SerpAPI with fallback scraping)
• Parses uploaded resumes (PDF/DOCX)
• Calls an AI (OpenRouter online or local Ollama) to generate an ATS-optimized LaTeX resume
• Compiles/Previews LaTeX on the client using latex.js and previews via react-pdf (subject to local tool compatibility)

Quick setup

1. Install dependencies:

cd d:/cv-generator; npm install

2. Create .env from .env.example and set keys:

• SERPAPI_KEY — required for SerpAPI scraping
• OPENROUTER_API_KEY — optional; if present, the app will use OpenRouter online
• OLLAMA_MODEL — optional; used when Ollama is running locally

OpenRouter mode

• Add OPENROUTER_API_KEY to .env. The serverless /api/generate route will call OpenRouter at https://openrouter.ai/api/v1/chat/completions.

Ollama local mode

• Run Ollama locally (example): ollama run or follow Ollama docs. The app will call http://localhost:11434/api/generate if OPENROUTER_API_KEY is not set.

SerpAPI usage

• Provide SERPAPI_KEY in .env. The /api/scrape route attempts SerpAPI first, then falls back to direct scraping.

LaTeX → PDF rendering

• The app generates LaTeX code. To create PDFs, compile the downloaded .tex file with a LaTeX distribution like TeX Live or MiKTeX.
• Client-side PDF preview is not implemented due to compatibility issues; use the LaTeX file with external tools.

Run the app

npm run dev
# then open http://localhost:3000 (or 3001 if port conflict)

Notes

• This scaffold uses base64 file upload for serverless API routes.
• You may want to secure the SerpAPI and OpenRouter keys in production.
• The AI prompt is in lib/aiGenerator.js. Tweak instructions for tone, length, or formatting.
• PDF preview requires server-side LaTeX compilation (not implemented here).
• If no AI backend is configured, the app returns a sample LaTeX for testing.