CVitae Studio

Demo Video

Context

A full-stack Next.js 16 platform that runs every AI feature on the Anthropic Claude API (Haiku 4.5), backed by MongoDB Atlas, Auth.js (GitHub and Google sign-in) and Upstash QStash for scheduled background jobs — deployed on Vercel and covered by 3,195 automated tests.

Problem

Job seekers need résumés that pass Applicant Tracking System (ATS) filters and still look polished to human eyes — then somewhere to track every version they send out, and hours every week to find openings worth applying to. Most tools solve one of these and ignore the rest, with little flexibility in export format.

Solution

CVitae Studio authors career data once against the JSON Resume standard, fully decoupled from design, so switching between five visual templates never touches the underlying facts. Every AI draft — bullet points, summaries, ATS rewrites and cover letters — runs through a Generate → Critique → Refine pipeline, and a hallucination guard flags any invented number or skill for explicit approval. Paste a job description (or import it from a posting link) for a 0–100 ATS score across four vectors, with synonym-aware matching and one-click keyword fixes, then export a designed or ATS-safe PDF, or a DOCX. Every application lives in an Airtable-style supertable with table and Kanban views, custom columns and a per-row activity log. Job Search Automation closes the loop: define a search profile and rules once, and a daily scan finds new openings, scores them against your résumé and drafts a tailored résumé and cover letter for your review — within daily spend caps.

Screenshots

Tech Stack