cv-boilerplate

A giveaway template for a recruiter-facing CV one-pager with a build-time PDF. A bilingual-capable storefront whose single job is to get a recruiter to download your CV PDF. The web page and the PDF render from one content source and never drift.

You fill in three things — your content, your config, optionally your theme — and deploy. Everything else is the reusable frame.

Created by Oleksandr Bilostotskyi. Licensed under Apache-2.0 — you may use it for your own CV and redistribute it, but you must keep the LICENSE and NOTICE attribution in any redistributed copy of the template itself. See NOTICE.

How it works

The site ships two renderings of the same dataset at different depths:

• CV page — the web one-pager (summary, one-line experience entries, a curated stack, contacts, languages). Its job is the PDF download.
• CV PDF — the "real CV" a recruiter submits to an ATS. A superset of the page: per-entry highlights and stacks, full skills, education, languages. No photo, structured for ATS parseability.

Both read from src/content/<locale>.ts. The PDF is generated at build time by scripts/build-pdf.tsx using @react-pdf/renderer — never edited by hand (ADR-0001). A parity check fails the build if locales drift structurally. Printing the page (Ctrl+P) is a separate, shallower path that appends a link to the full PDF.

Routing is config-driven (ADR-0002): the default locale renders at /, additional ones at /<code>. Ships with English only.

Discoverability artifacts — robots.txt, sitemap.xml, the JSON-LD graph, the per-locale OG cards and llms.txt (an llmstxt.org map for AI crawlers) — all derive from the same config and content, so they never drift from the CV. llms.txt renders the default locale.

Fill it in (the 3 surfaces you own)

Read CONTEXT.md for the vocabulary first. Then:

1. Content — src/content/en.ts

Your career data. Every field ships as a structurally-valid placeholder ("Your Name", one example role) so the build works before you touch it. Replace the values; keep the shape (enforced by src/content/types.ts). The labels block is UI chrome (section titles, button text), not personal data.

Sections self-hide when empty (page and PDF): empty an array (experience: [], projects: [], skills: [], pageStack: [], languages: []), set summary: '', or drop the education key — that section simply does not render. name / title / location are always shown. So a newcomer with no jobs fills projects (pet projects, open-source, hackathons — each with an optional url to the repo/demo) and leaves experience: []; the "Work Experience" heading never appears.

2. Config — src/shared/site.ts

Your non-career settings: url, email, pdfBaseName, socials, address, optional selectedWork footer link, and the localeConfigs list.

3. Theme (optional) — src/shared/theme.ts + src/app/globals.css

Colors and fonts. The JS palette (brand) in theme.ts feeds the PDF and OG card; the @theme block in globals.css is its mirror for the web page. Change both, in sync — see DESIGN.md.

Adding a language

1. Append an entry to localeConfigs in src/shared/site.ts with four fields: code (ISO 639-1), label (the toggle text — a language name, never a country: УКР, not UK, which reads as Britain), ogLocale (e.g. uk_UA), and ogNameLetterSpacing (optical tightening of the name on the OG card — denser scripts like Cyrillic read heavier, so use a tighter value there). The first entry is the default and renders at /; others at /<code>.
2. Add src/content/<code>.ts and register it in src/content/index.ts.
3. Keep locales structurally parallel — the parity check compares language-independent facts (entry counts, company names, page stack).
4. If the script needs extra font subsets (e.g. Cyrillic), add them in src/shared/fonts.ts and confirm the committed TTFs in assets/fonts cover the glyphs (they carry Latin + Cyrillic by default).

Routing, metadata, sitemap, the parity loop, the language toggle and the per-locale OG cards need nothing more — they all iterate localeConfigs. The only default-locale artifact is llms.txt (the llmstxt.org spec expects a single /llms.txt at the site root).

Tech stack

• Next.js 16 (App Router, static export) + React 19
• Tailwind CSS 4
• TypeScript 5
• @react-pdf/renderer 4 — build-time PDF generation
• @vercel/analytics + speed-insights
• pnpm

Getting started

Requires Node.js 20+ and pnpm.

pnpm install
pnpm dev          # http://localhost:3000

Scripts

Command	Description
pnpm dev	Start the Next.js dev server
pnpm build	Production build (runs prebuild → PDF generation first)
pnpm pdf	Regenerate the CV PDF(s) from content
pnpm lint	ESLint
pnpm format	Prettier write
pnpm typecheck	tsc --noEmit

pnpm build runs prebuild automatically, which validates content parity and writes the PDF(s) (named from pdfBaseName) into public/.

Project layout

src/
  app/              Next.js App Router — [[...locale]] route, robots, sitemap, llms.txt
  components/       CV page components, download button, locale toggle, JSON-LD
  content/          Content source: en.ts (placeholders), index.ts, types.ts
  shared/           site config, theme tokens, metadata, fonts, parity, profile
scripts/            build-pdf.tsx + pdf-document.tsx (react-pdf layout)
assets/fonts/       Manrope + JetBrains Mono TTFs (OFL — for PDF/OG glyphs)
docs/adr/           Architecture decision records

Deployment

Deployed as static HTML (output: 'export'). Every content edit regenerates the PDF on build, so the download link can never serve a stale CV.

License

Apache-2.0 © 2026 Oleksandr Bilostotskyi. Attribution required — see NOTICE. The attribution covers the template; the CV content you author on top of it is yours.