The Brief
A promising job title alone does not tell you which experience in your CV addresses the listing, or which requirements you have not demonstrated. CV Job Matcher was built to compare a candidate's own text to a particular real vacancy and show actionable feedback for that vacancy. It is distinct from JobAI Web's multi-feed discovery search and from the AI Smart Job Finder Android app.
Real Listings First
A Next.js API searches Remote OK's public remote-job feed. Adzuna results are included only when its optional API credentials are configured; a missing provider simply yields fewer results. Each result shows its source, company, location and an outbound original-listing link. Jobs are not invented to support an AI score.
Match One Job at a Time
Visitors can paste CV text or extract text from a PDF, DOCX or TXT file in the browser, then choose Analyze match on a listing. The server sends that CV text together with the selected job description to OpenRouter for an AI-generated score, strengths, gaps and concrete suggested CV edits. A primary model and a free-model fallback are tried, and incomplete or invalid AI output is rejected instead of displayed as a confident result. Scores are model-generated guidance, not a verified hiring prediction.
Privacy and Architecture
The file itself is parsed in the browser; analyzing a match still transmits the extracted CV text to the app's API and to the external AI provider. The app does not implement persistent CV or search history: results live in browser memory for the session. Its Next.js API routes run via the Cloudflare Workers adapter, with model credentials kept on the server. Avoid submitting sensitive personal data to a public demo.
The Outcome
A live demonstration of a bounded AI feature: match the user's CV against an actual listing, explain the result with job-specific details and provide a link back to the source. Feed availability and model quotas can affect the output; neither an AI score nor a job listing guarantees an interview or an open vacancy.