Generate ATS-Optimized CVs from Multiple Sources

CV Generator turns raw profile data - LinkedIn, GitHub, a portfolio site, an existing draft, or a step-by-step questionnaire - into a polished, ATS-ready CV, outputting a paste-ready plain-text version formatted specifically for FlowCV, Canva, Google Docs, or Word, plus a scored flaw report and a missing-information checklist. It's an explicitly versioned rewrite that documents and fixes 15 flaws from prior drafts: plain-text-first output instead of Markdown, real FlowCV/Canva field mapping, one questionnaire question at a time instead of a 20-question dump, a structural anti-hallucination enforcement gate before every output, a separately-scoped cover letter block, a 0-100 scored ATS flaw report with per-item pass/fail, a stated mid-level default when seniority can't be detected, an explicit field-by-field paste map for what FlowCV/Canva can't render, a hard tense-consistency gate that blocks output until corrected, a machine-checkable banned-phrase blocklist, and a hard fallback to requesting a LinkedIn PDF export when scraping is blocked rather than proceeding on empty data.

Use it to generate a professional CV from multiple merged sources, tailor an existing CV to a specific job description, improve the language/metrics/structure of a draft resume, or produce a paste-ready version for FlowCV or Canva specifically. Any combination of sources (LinkedIn URL or PDF, portfolio site, questionnaire, existing CV/draft, GitHub profile, resume file) is valid and gets merged and deduplicated before writing; if no source is given, it defaults immediately to the questionnaire. If a LinkedIn page fetch is blocked or empty, it stops immediately and asks for a PDF export rather than generating from nothing.

After source selection, it asks the CV's purpose (specific job application, general professional CV, internship/entry-level, academic/research, freelance proposal, career change, executive, military-to-civilian, or return-to-work), detects seniority from the data (defaulting to mid-level 3-8 years with an explicit stated assumption if undetectable), and picks a format (chronological default, functional/skills-first for career changers or gaps, hybrid, academic, executive/board, or portfolio-led for creatives). Extraction rules are source-specific: LinkedIn URL extraction pulls name/headline/contact/about/experience/education/skills/certifications/projects/achievements/volunteer/languages/publications in order; LinkedIn PDF and resume files get OCR with an accuracy warning if scanned; GitHub extraction covers pinned repos, bio, tech stack, and contribution summary. Writing rules enforce STAR-lite experience bullets, a machine-checkable banned-phrase blocklist (no "passionate about" and similar filler), and strict tense consistency as a hard gate. ATS optimization covers structural rules, keyword strategy tied to job-description integration, and platform-specific notes. Final output is delivered in a fixed order: a staging draft, a FlowCV paste-ready version, a Canva paste-ready version, a missing-information checklist, a CV flaw report scored 0-100 with per-item pass/fail, 3-7 specific improvement suggestions, and a suggested file name - with an optional separate cover-letter block.

Country and market conventions are documented explicitly (including Nepal/South Asia specifics), and a decision tree ties source, purpose, seniority, and format selection together into one consistent flow.

Job seekers and career coaches who want a rigorously ATS-optimized, fact-checked CV built from real profile data and pasteable directly into FlowCV or Canva, rather than a generic AI-drafted resume with invented details or unverified claims.