As an engineer, I was incredibly frustrated by the modern resume builder market. Every tool either had a terrible UI, demanded a monthly subscription right before I clicked "Download", or used outdated text-parsers that completely broke when I tried to import my old resume.
So, I decided to build my own:
ZenResume
I wanted to challenge myself to make it as fast and lightweight as possible, so I skipped React and Next.js entirely. The entire application is built using Vanilla JavaScript (ES6) and Firebase for authentication.
The Coolest Part: Gemini Vision API The biggest challenge with resume builders is extracting data from old PDFs. Traditional parsers fail if the PDF is image-based. I solved this by integrating Google's Gemini 2.5 Flash Vision API. When you upload a resume, ZenResume takes a picture of it, sends it to Gemini with a strict JSON schema, and Gemini visually extracts your history and maps it perfectly into my ATS-compliant templates.
Handling AI Rate Limits Because Gemini's free tier only allows 15 requests per minute, I implemented a Client-Side Exponential Backoff algorithm. If the AI is busy, the fetch request silently catches the 429 error, waits 3 seconds, and retries automatically without crashing the UI!
It's 100% Free I built this to help the community, especially freshers looking for tech jobs. You can export high-quality A4 PDFs without watermarks.
I would absolutely love your feedback on the UI and the AI extraction accuracy! Let me know what you think:
ZenResume
Top comments (0)