P2Pool Starter Stack — organization website

The marketing / showcase site for the P2Pool Starter Stack — an orchestrator (Pithead) and the miners that feed it (RigForge) — plus the node-starter stacks to come.

Built with Hugo (extended). Industrial-cypherpunk look, dark + #F26822, zero third-party requests (system fonts only — nothing phones home, in keeping with the privacy ethos).

Run it locally

You need Hugo extended (≥ 0.162). On macOS: brew install hugo.

hugo server          # live reload at http://localhost:1313

CI and the deploy build pin an exact Hugo version in .hugoversion — a single source of truth read by both ci.yml and deploy.yml, so they can't drift. Bump it by editing that one file (a weekly hugo-bump.yml job also opens a PR when a newer Hugo ships).

Build

hugo --gc --minify   # static output to ./public

Test

CI (.github/workflows/ci.yml) runs on every push/PR and is two checks:

# 1. Strict build — any Hugo warning (deprecation, missing data, …) fails the build
hugo --gc --minify --panicOnWarning --baseURL "http://localhost/"

# 2. Link / anchor / image check on the built site (internal only, deterministic)
#    https://github.com/wjdp/htmltest  — config in .htmltest.yml
htmltest

Both must pass before main deploys. External links are skipped by default so CI never flakes on a third-party outage — run htmltest with CheckExternal: true to audit them.

Where the content lives

All landing-page copy is in data/content.yaml — edit text there, not in the templates. The page is assembled by layouts/index.html from that data.

The landing page introduces both Pithead OS and Docker Compose for new users, alongside mining, privacy, dashboard, fleet, and recovery features. Keep optional/beta defaults explicit and link to the install guides and release notes for detailed requirements and known issues. The roadmap summarizes shipped capabilities; release numbers stay in the project badges. Pithead's JSON-LD description reuses its project blurb so search metadata stays aligned with the visible copy. The hero pairs an appliance image/checksum download with the DIY setup terminal. Image URLs use the Pithead tag in data/releases.json; keep the versioned release asset names in sync if they change.

data/content.yaml        ← all copy (hero, stats, projects, roadmap, ethos, FAQ, CTA)
data/releases.json       ← latest pithead / rigforge release tags (version badges)
layouts/index.html       ← single-page assembly
layouts/partials/        ← head / header / footer / icon set / schema (JSON-LD)
assets/css/main.css      ← the entire visual system
assets/js/main.js        ← reveal-on-scroll, header state, copy-to-clipboard
scripts/refresh-releases.py  ← refreshes data/releases.json from the GitHub API
static/img/              ← marks + dashboard screenshot + og-card
static/favicon.svg       ← P2Pool Starter Stack layered mark (modern browsers)
static/favicon.ico       ← raster fallback (Safari/legacy tabs) — 16/32/48
static/apple-touch-icon.png  ← iOS / Safari home-screen icon (180×180)
static/robots.txt        ← crawl rules (search + AI assistants explicitly welcome)

The icon/mark set is inline SVG in layouts/partials/icon.html ({{ partial "icon.html" "pithead" }}), so every mark inherits color from CSS via currentColor.

Discoverability (SEO / GEO)

The site is tuned to be found — and cited — when people (or their AI assistants) ask how to mine Monero privately. The copy is keyword-aware without being spammy, and three machine-readable layers back it up:

• layouts/partials/schema.html emits one application/ld+json graph: an Organization, a WebSite, a SoftwareApplication for each of Pithead and RigForge (free, MIT, OS-tagged), and a FAQPage. The FAQ entries are generated from the same data/content.yaml → faq block the page renders, so the visible Q&A and the structured data never drift. Built through jsonify | safeJS, so escaping is always valid.
• layouts/partials/head.html carries the title/description, canonical, Open Graph + Twitter card (static/img/og-card.png), and robots: max-image-preview:large.
• static/robots.txt allows everything and explicitly welcomes answer-engine crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, …) plus the auto-generated sitemap.xml.

After editing the FAQ or any project copy, paste the built page into Google's Rich Results Test to confirm the JSON-LD still validates.

Version badges (kept current automatically)

Each project card shows the latest release tag of its repo (e.g. v1.0.2). To keep the site's zero third-party requests promise, the version is baked in at build time — the page never calls out at runtime, and CI stays deterministic (no network):

• The Hugo build only reads data/releases.json — a committed fallback that the page renders from.
• scripts/refresh-releases.py updates that file from the GitHub Releases API. It's best-effort: on any API/network/rate-limit error it keeps the committed value, so it can never fail a build. Set GH_TOKEN to dodge the unauthenticated rate limit.
• The deploy workflow runs that script before each Hugo build (with the built-in GITHUB_TOKEN) and on a daily schedule, so the live badges track new releases without a manual push.

Refresh the committed fallback locally with python3 scripts/refresh-releases.py.

Deploy

This repo is the org's GitHub Pages site (p2pool-starter-stack.github.io), served at https://pitheadmining.com/. The old github.io address redirects there.

.github/workflows/deploy.yml builds with Hugo and publishes on every push to main. In the repo's Settings → Pages, set Source: GitHub Actions once and it deploys automatically thereafter.

To put it on a custom domain later: add static/CNAME containing the domain and set baseURL in hugo.toml.

License & contributing

Released under the MIT License — same as Pithead and RigForge.

Contributions are welcome — see CONTRIBUTING.md.