jeremy muise

Personal Homepage (Hugo)

This project is a Hugo-powered personal homepage with automated GitHub Pages deployment via GitHub Actions.

1) Configure site settings

Edit hugo.yaml:

If the repository is named <your-username>.github.io, use:

2) Customize homepage content

Edit:

3) The PDF résumé

The “Résumé” page renders twice from the same content/resume.md and data/resume_skills.yaml: once as the styled web page, and once through a second Hugo output format (PRINT, configured in hugo.yaml) that produces a plain, black-on-white, single-column HTML document with no nav/icons/decorative chrome — see layouts/_default/single.print.html. That plain document exists so the PDF built from it survives ATS résumé parsers: no Nerd Font icon glyphs in the text layer, no navigation text ahead of your name, contact info as plain visible text, and the full (untiered) skills list as comma-separated text rather than badges. CI (.github/workflows/hugo.yml) renders that page to public/resume.pdf with headless Chromium (scripts/render-resume-pdf.mjs, via Puppeteer) after the Hugo build, so it ships as part of the same Pages deploy — no separate hosting or repo needed.

To customize what’s on the PDF:

params.resumePdfUrl (already set to resume.pdf) is what the web page links to for the download; it doesn’t need to change unless you rename the output file.

To generate the PDF locally: hugo (builds public/print/resume/index.html among the rest), then npm ci once, then node scripts/render-resume-pdf.mjs public/print/resume/index.html public/resume.pdf. Headless Chromium needs a handful of system shared libraries not installed by default in a minimal container (libnspr4, libnss3, libcairo2, libasound2t64, and similar) — install those via your OS package manager if puppeteer.launch() fails with a “cannot open shared object file” error; GitHub Actions’ ubuntu-latest runners already have them.

4) Enable the contact form

GitHub Pages serves static files only, so there is no server to send mail from. The contact page posts to Formspree, which relays each submission to your inbox.

  1. Sign up at formspree.io and create a new form.
  2. Copy the form ID — the part after /f/ in the endpoint it gives you (https://formspree.io/f/abcdwxyz → abcdwxyz).

  3. Put it in hugo.yaml:

    params:
  contact:
    formspreeID: "abcdwxyz"
  4. Confirm the verification email Formspree sends, then submit the live form once to check it arrives.

Until formspreeID is set, the contact page renders a short “not configured yet” notice instead of the form. The form ID is not a secret — it is a public endpoint, safe to commit.

Notes:

5) Enable visitor analytics (optional)

Two independent, optional trackers, both configured under params.analytics in hugo.yaml. Leave either blank to skip it — nothing renders until it’s configured.

Page-view analytics — GoatCounter

Free, cookieless, no consent banner required.

  1. Sign up at goatcounter.com and pick a site code (yours becomes https://<code>.goatcounter.com).

  2. Put just the code in hugo.yaml:

    params:
  analytics:
    goatcounterCode: "your-code"
  3. View stats at https://<code>.goatcounter.com.

Visiting-company identification — Leadfeeder

Flags which organization a visit came from when it’s identifiable from a corporate network (home/mobile/VPN traffic won’t be identified — that’s a limit of the underlying IP-to-org data, not a bug). Free “Lite” plan covers the last 100 identified companies/month with 7-day history — plenty for a personal site.

  1. Sign up for the free plan at leadfeeder.com and add your site.
  2. In the dashboard: gear icon → Company → Website Tracker → Tracker Script, then copy the whole script it gives you (it’s generated per-account, not a fixed snippet).

  3. Paste the entire thing into hugo.yaml as a block scalar:

    params:
  analytics:
    leadfeederScript: |
      <script>...the script Leadfeeder gave you...</script>
  4. Set up email alerts inside Leadfeeder’s own dashboard, under notification settings — that part isn’t controlled from this repo.

A privacy note: GoatCounter needs no disclosure since it doesn’t process personal data. Leadfeeder-style identification does derive information from visitor IP addresses, so if you’re aiming for strict GDPR compliance you may want a short privacy note on the site; for a low-traffic personal portfolio the practical risk is minimal, but it’s your call to make.

6) Publish on GitHub Pages

  1. Create a GitHub repo and push this project to the main branch.
  2. In GitHub repo settings, open Pages.
  3. Set Source to GitHub Actions.
  4. Push to main again (or use Run workflow in Actions).
  5. Your site will be deployed automatically.

7) Local preview (optional)

For day-to-day content editing, install Hugo and run:

hugo server -D

Open http://localhost:1313. This is a live-reloading dev server that renders everything in memory — it never touches public/ and never runs the PDF-render script, so the “download the detailed PDF résumé” link will 404 here. That’s expected, not a bug.

To preview the site as it will actually deploy — including a working PDF link — build it for real and serve the static output instead:

npm ci    # once
npm run preview

This runs hugo --gc (with whatever baseURL is set in hugo.yaml — the real production one), renders public/resume.pdf with Puppeteer, and serves public/ at http://localhost:3000. Because baseURL is the bare https://jmuise.github.io/ with no repo-name subpath, the site’s internal links are already root-relative (/css/site.css, not /some-repo/css/...), so they resolve the same way locally as they will in production — no baseURL override needed for local serving.

npm run build does the same build + PDF render without serving anything — useful if you just want the files on disk (it’s also what CI runs).

Project structure