How AI Job Answers Works
Three steps from your CV to a tailored cover letter, application answer, or CV review. The whole flow runs in the browser plus one round-trip to our generation API — no signup, no storage, no waiting.
- 1
Upload your CV
Drag in a PDF or DOCX. Parsing happens in your browser using pdf.js and mammoth.js — the file itself never leaves your device. If parsing fails (some scanned PDFs have no extractable text), there is a paste-text fallback. Your parsed CV is saved in localStorage so you only do this once.
- 2
Paste the job description
Copy the JD from the company's careers page, LinkedIn, or wherever you found it. Anything from 100 characters up. The JD lives in the browser session only — closing the tab clears it.
- 3
Generate
Pick a tab — Cover Letter, Application Question, or CV Evaluation. Pick a length (where applicable) and hit Generate. The output streams in real-time as the model writes, so you can see it forming. Edit it inline, copy with one click, paste into the actual application form.
What happens behind the scenes
Generation runs on a Cloudflare Pages Function that calls Gemini 2.5 Flash. The prompt includes your CV text, the job description, and a target length — nothing else. There is no persistent context, no memory of previous sessions, and the request body is not logged.
Cloudflare Turnstile sits between the form and the generation endpoint to block scripted abuse, and per-IP rate limits cap usage so a viral moment doesn't make this expensive enough to take offline.