About

Most CV tools send your document directly to an AI API. Name, phone number, address, employment history - transmitted unredacted. The model that rewrites your resume has full access to your identity. We treat that as a design failure, not an acceptable tradeoff.

Before any outbound request, a dedicated sanitization step removes every personal identifier from your CV. The three models working on your document - one for generation, one for hiring-manager critique, one for authenticity detection - receive only anonymized professional content. Your real identifiers are reinserted locally in your browser after the response returns. Your identity is never part of the AI processing chain.

The pipeline is structured to prevent the most common failure in AI writing tools: a model grading its own output. Generation and critique are handled by entirely separate models with no shared context. The critique step works against 10 named failure categories - unsupported claim, seniority mismatch, generic language, weak opening, low evidence density, and five others. A third model then runs a 7-dimension authenticity check, scoring each dimension from 0 to 1. Output only proceeds if all dimensions pass. If anything fails, a revision runs against the structured failure objects - not a vague re-prompt.

Every claim in your rewritten CV can be traced back to what you actually documented. A dedicated explanation step maps each bullet and theme to its source section in your original CV. If a claim cannot be grounded in what you submitted, it is not included. No fabricated metrics. No inflated scope. No seniority claims beyond what your dates and roles support.

What you receive: a rewritten CV with a full change log of every edit and the reason for it; an ATS compatibility score with specific gaps and keyword mismatches identified; three cover letter styles - each structurally distinct, each passing the authenticity checklist; and interview prep Q&A anchored to the role and your background.

Built for professionals sending more than one application. Pay once per optimization. No subscription, no auto-renewal, no card stored until you choose to proceed. Under 30 seconds for the full pipeline.

When processing completes, the uploaded file, the extracted text, and the job description are gone. No file storage. No database entry. No reconstruction possible. What is retained is the generated output - accessible only to your account.