Skip to content
BranchesTags

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
.dockerignore
.dockerignore
 
 
.gitignore
.gitignore
 
 
LICENSE
LICENSE
 
 
README.md
README.md
 
 
config_example.json
config_example.json
 
 
go.mod
go.mod
 
 
go.sum
go.sum
 
 
main.go
main.go
 
 
main_test.go
main_test.go
 
 
sheetproxy
sheetproxy
 
 

Repository files navigation

sheetproxy

Static mirrors of Google Sheets

Environment Variables

Shared settings (used by every sheet):

Variable Required Default Description
GIT_PAT no GitHub PAT for push (shared across all sheets)
GIT_EMAIL no sheetproxy@local Git commit email
GIT_NAME no sheetproxy Git commit name
POLL_MINUTES no 10 Default regeneration interval
CONCURRENCY no 8 Concurrent fetches
CONFIG_FILE no /srv/sheetproxy/config.json Path to a JSON config file (see below)
CWEBP_BIN no cwebp Path to the cwebp binary used for WebP encoding

Images: lossless WebP

Images are always stored losslessly. Large images and any JPEG source are re-encoded to lossless WebP (falling back to PNG), so the proxy never emits a lossy JPEG. This requires the cwebp binary at runtime (install it locally, e.g. apt install webp / brew install webp). Set CWEBP_BIN if it is not on PATH.

Single sheet (backward compatible)

Variable Required Default Description
SHEET_URL yes* Google Sheet URL
WWW_DIR no ./www Output directory
GIT_REPO no GitHub repo (e.g. user/repo)
PAGE_TITLE no Frank Tracker Page title
PAGE_DESCRIPTION no, auto auto Meta description (SEO)
ARTIST_NAME no, auto auto Artist name, derived from page_title
PAGE_LANG no en <html lang> attribute
FAVICON_URL no Favicon URL
ANALYTICS no true Enable Plausible analytics
PLAUSIBLE_SCRIPT no Plausible script URL

* SHEET_URL is required only when no CONFIG_FILE is set (including the default path). If CONFIG_FILE is explicitly set but the file is missing, the process errors out.

Multiple sheets

Set CONFIG_FILE to a JSON file, or rely on the default /srv/sheetproxy/config.json (used automatically when CONFIG_FILE is unset). The top level holds shared/global settings (including the GitHub git_pat, shared by every sheet), and a jobs array lists each sheet. Sheets are identified by sheet_id (the ID from the spreadsheet URL), or sheet_url as a fallback. See config_example.json for a full example:

{
  "git_pat": "ghp_xxx",
  "git_email": "sheetproxy@local",
  "git_name": "sheetproxy",
  "poll_minutes": 10,
  "concurrency": 8,
  "jobs": [
    {
      "sheet_id": "1A2b3C4d...",
      "www_dir": "./www1",
      "git_repo": "user/repo-one",
      "page_title": "Tracker One",
      "page_description": "Release tracker for Tracker One"
    },
    {
      "sheet_id": "2B3c4D5e...",
      "www_dir": "./www2",
      "git_repo": "user/repo-two",
      "page_title": "Tracker Two",
      "favicon_url": "https://example.com/fav.ico",
      "analytics": false
    }
  ]
}

SEO

Each generated page is injected with auto-generated, domain-free SEO and AI/answer-engine (GEO) friendly markup. No site domain is required or hardcoded; everything is derived from the sheet and page_title.

  • Basic: <title>, <html lang>, meta description, keywords, author, application-name, format-detection, theme-color, color-scheme.
  • Crawler / AI directives: robots, googlebot and bingbot set to index, follow, max-image-preview:large, max-snippet:-1, max-video-preview:-1 so bots and answer engines can use full snippets and large images.
  • Mobile / PWA: apple-mobile-web-app-*, mobile-web-app-capable.
  • Open Graph: og:title, og:description, og:type, og:locale, og:site_name, and og:image (reused from the sheet's own preview image) plus og:image:alt and og:image:type.
  • Twitter Card: summary_large_image with twitter:title, twitter:description, twitter:image and twitter:image:alt.
  • Structured data: a JSON-LD @graph linking MusicGroup, Person, WebPage, WebSite, Organization, Dataset, BreadcrumbList and FAQPage nodes — cross-referenced via @id with member/memberOf, mainEntity, isPartOf, publisher, provider, creator, about, subjectOf, logo, primaryImageOfPage and isFamilyFriendly.
  • No visible changes: everything is injected into <head> or emitted as separate files — the rendered page body is untouched.
  • Freshness: LastUpdate is stamped into og:updated_time, article:published_time/modified_time, dcterms.date/dcterms.modified and dateModified in the structured data on every regenerate.
  • Dublin Core metadata: dcterms.title, creator, subject, description, language, type, date, modified plus copyright.
  • Performance hints: preconnect/dns-prefetch for the analytics host, gstatic and docs.google.com.
  • Mobile/OS: dir="ltr", light/dark theme-color, msapplication-TileColor/TileImage, handheldfriendly, MobileOptimized, apple-mobile-web-app-* and mobile-web-app-capable.
  • Sharing: a dedicated og-card.png (1200x630, cover-cropped from the sheet preview) is generated and used for og:image/twitter:image, with accurate width/height/type; twitter:label1/data1 + label2/data2.
  • Images: sheet images get loading="lazy", decoding="async", fetchpriority="low", crossorigin and referrerpolicy="no-referrer".
  • AEO / LLM: llms.txt (and a fuller llms-full.txt) are generated with the site summary and key topics, and robots.txt explicitly allows AI/answer-engine crawlers (GPTBot, ChatGPT-User, Google-Extended, ClaudeBot, anthropic-ai, PerplexityBot, CCBot) plus standard crawlers.

PWA

Each site is installable as a PWA. A manifest.webmanifest is generated with id, name, short_name, description, start_url, scope, standalone display (with display_override), dark theme_color/background_color, lang and icons: icon-180.png/icon-192.png/icon-512.png plus a icon-maskable-512.png (all resized from the sheet's preview image), the favicon and the preview image itself. The <head> links the manifest and an apple-touch-icon (including a 180x180 variant) and includes apple-mobile-web-app-* / mobile-web-app-capable metas so it can be added to a home screen on iOS and Android.

The artist name is derived from page_title (e.g. Frank Ocean Tracker → Frank Ocean; override with artist_name) and used to generate the description ("View and explore {artist}'s unreleased and released discography and music leaks on {title}."), a leak-heavy keywords list (unreleased music, music leaks, leaks, grails) and the structured data. Set page_description to override the generated description.

A canonical link / og:url are deliberately not emitted since they require an absolute site domain.

Each sheet runs in its own goroutine, generating into its own www_dir and force-pushing a single squashed commit to its own git_repo.

Global JSON settings are overridden by the equivalent environment variables when both are present.

Build & Run

CGO_ENABLED=0 GOOS=linux GOARCH=amd64 go build -trimpath -ldflags="-s -w" -o sheetproxy main.go
./sheetproxy

About

No description, website, or topics provided.

Resources

Readme
GPL-3.0 license
Activity
Custom properties

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

Packages

Contributors

Languages