3 3.1 ready-for-dev 2025-12-01 BMAD Story Context Workflow docs/sprint-artifacts/3-1-create-pdf-generation-api-route.md system a server endpoint that generates PDF from the resume page users get consistent, high-quality PDF output - Create API route file at server/api/resume/pdf.get.ts - Implement Puppeteer PDF generation - Add error handling with timeout - Configure for Vercel deployment - Test API endpoint Given a GET request to /api/resume/pdf, when the server processes the request, then it returns a PDF binary with Content-Type application/pdf Response includes Content-Disposition: attachment; filename="Ali_Arghyani_Resume.pdf" PDF matches the web preview exactly (WYSIWYG) PDF text is selectable and copy-able (ATS-compatible) PDF is A4 format (210mm × 297mm) PDF generation completes in under 3 seconds Given an error occurs, when caught, then it returns status 500 with JSON error message Timeout is set to 10 seconds max docs/architecture.md

API Contracts

GET /api/resume/pdf returns PDF binary with Content-Type: application/pdf and Content-Disposition header docs/architecture.md

Novel Pattern: WYSIWYG PDF Export

Puppeteer navigates to /resume?print=true, waits for networkidle0, generates PDF with format A4 and printBackground true docs/architecture.md

Deployment Architecture

Use puppeteer-core + @sparticuz/chromium for Vercel serverless. Memory: 1024MB, maxDuration: 10s docs/sprint-artifacts/tech-spec-epic-3.md

APIs and Interfaces

GET /api/resume/pdf - Success returns PDF buffer, Error returns 500 with JSON { error, message } app/pages/resume.vue Resume page that will be captured by Puppeteer. Supports ?print=true query param. isPrintMode computed property app/composables/useResumeData.ts Provides getPdfFilename() for generating filename getPdfFilename() - File location must be server/api/resume/pdf.get.ts (Nuxt server route convention) - Must use defineEventHandler from Nuxt - Must detect environment for puppeteer vs puppeteer-core selection - Must navigate to /resume?print=true (not /resume) - Must wait for networkidle0 before PDF generation - Must close browser in finally block to prevent memory leaks - Timeout must be 10 seconds max - Memory limit 1024MB on Vercel defineEventHandler Nuxt server utility defineEventHandler(async (event) => { ... }) nitro/runtime setResponseHeaders Nuxt server utility setResponseHeaders(event, { 'Content-Type': string, 'Content-Disposition': string }) h3 getRequestURL Nuxt server utility getRequestURL(event): URL h3 puppeteer.launch Puppeteer API puppeteer.launch({ headless: boolean, args?: string[], executablePath?: string }): Promise<Browser> puppeteer or puppeteer-core page.pdf Puppeteer API page.pdf({ format: 'A4', printBackground: boolean, margin?: object }): Promise<Buffer> puppeteer Nuxt server route testing. Test API response headers and PDF content. server/api/**/*.spec.ts, tests/ Request /api/resume/pdf, verify Content-Type is application/pdf Verify Content-Disposition header contains correct filename Open generated PDF, compare visually to web preview Open PDF in reader, try to select and copy text Check PDF page dimensions are A4 (210mm × 297mm) Measure time from request to response, verify under 3 seconds Simulate error, verify 500 status and JSON response Verify timeout configuration in code