CVera

Candidates rarely know which technical signals a job post actually expects from their résumé.

A candidate should see which skills from a software job post their résumé contains or lacks. The document should not be sent to a third-party analysis provider, and the score should not be presented as a hiring decision.

CVera validates PDF/DOCX signatures, size and readable text, parses the file in request memory, then compares detected skills with a Turkish or English software job post using traceable rules. Matches and gaps are reported separately.

Next.js 16 and React 19; pdf-parse and mammoth extract text, server-side validation checks file signature, size and résumé signals before analysis.

Parsing in request memory and sending no-store responses avoids persistent file storage and response caching. Deterministic skill rules keep the preview explainable, but the score is not a commercial ATS or hiring decision.

The open-source MVP returns an explainable compatibility preview and separate matched/missing skill lists. GitHub Actions checks lint, unit tests and production builds; scanned PDFs without text are not yet supported.

An explainable keyword-and-structure score is more honest than a black-box number — as long as its limits are stated clearly.