JobVouch

- ATS scoring is not magic. It breaks down into keyword matching, skills coverage, experience relevance, and formatting. Making that transparent to users was the biggest unlock.

- Truthfulness matters. The AI never invents skills. It just reorganizes and rephrases what you already have. Users trust it because of this.

- Resume parsing from PDFs is harder than it sounds. Spent more time on that than on the AI itself.

Revenue: Freemium model. Free tier gives 3 tailored resumes per month. Paid plans from $2.99/week.

Not here to sell anything. Just sharing the build in case it resonates with anyone else grinding through job apps.

I was applying to jobs and got tired of the same cycle: find a posting, open my resume, spend 45 minutes swapping keywords, reordering bullets, tweaking my summary. Multiply that by 20+ applications a week. It was soul-crushing.

So I built JobVouch. You paste a job description, upload your resume (or build one in the app), and the AI rewrites your content to match what the employer wants. It scores your resume against the JD on a 0-100 scale and shows you exactly which keywords you are missing.

The stack: Next.js frontend, FastAPI backend, PostgreSQL, Redis for caching, AWS S3 for file storage. The AI tailoring runs through parallel processing with rate limiting to handle multiple bullet rewrites efficiently.