This is the source of https://survivejs.com/.
First, clone this repository:
git clone https://github.com/survivejs/website-v3Set up a .dev.vars file:
GET_IMAGES_ENDPOINT=TODO
GET_IMAGES_TOKEN=TODO
IMAGES_API_URL=TODO
IMAGES_ROOT=TODOBootstrap the project:
npm run bootstrapStart the development server:
npm run startNow you can open the local Wrangler URL.
Build using the following command:
npm run buildValidate the Cloudflare Worker bundle:
npm run worker:dry-runBook overviews show the latest commit affecting the book repository’s
manuscript directory. Chapters show the latest commit affecting their own
source file, following renames. Dates use the Git committer date and link to
the exact commit; they do not indicate a technical review of the material.
These dates come from the checked-out repositories under books/. Run
npm run fetch:book-repositories before building to refresh those checkouts.
Keep their full Git history: builds from source archives or shallow clones
show “Git history unavailable” instead of a guessed date.
Run local Lighthouse audits for mobile and desktop:
npm run lighthouse:localThe command builds the site, serves ./build locally, and writes HTML/JSON
reports to reports/lighthouse/.
If Chromium is missing, install the pinned browser first:
npm run playwright:installFor local Agent CI and more Lighthouse options, see
docs/development.md.
When book routes move, generate a Disqus URL Mapper CSV from a Disqus export:
node utils/generate-disqus-rewrites.js \
../site-disqus/survivejs-2026-05-16T09_44_20.025877-all.xml \
/tmp/disqus-url-map.csvUpload the generated CSV in Disqus under Discussions > Tools > Migrate Threads
Upload URL Map. Disqus documents this as the required path when a thread URL changes by more than the base domain.
The site content is available under CC BY-NC-ND license. So, as long as there's a proper attribution, you can reuse the content. Ideally, you would contribute your improvements back, but that's not necessary.
The site source is available under MIT license.
Page metadata is rendered centrally by site/utilities/metadata.ts. Every content
page must have a title and description; a missing value fails the build. Book
chapters derive these from the manuscript, and topic archives get a description
from their topic name. Canonical and Open Graph URLs use the production site URL
and match sitemap paths. The error page is noindex; the feed and verification
file are excluded from the sitemap.
JSON-LD describes pages, blog posts, the site, and the author profile. Blog dates
come from frontmatter; book modification dates come from Git, with no build-date
fallback. Atom entries use the same canonical URLs, authors, and source dates.
The shared 1200×630 social preview is bundled at /images/social.png; regenerate
it with npm run generate:social-image (requires Playwright Chromium).
Run npm test, npm run build:site:validate, then npm run check:metadata to
check the rendered metadata and sitemap together.