The full upstream README, mirrored here for reference. Install config, tool schemas, adoption signals, and an original overview live on the Jedanderson Corpus MCP listing page.
Personal corpus of Jed Anderson — essays, papers, and concept pieces on Environmental Superintelligence, information physics, and related first-principles inquiries.
Live site: https://jedanderson.org
Content in src/content/ is licensed CC-BY-4.0 unless otherwise noted in the license: field of an individual file's frontmatter. Some foundational treatises are released under CC0 — check the frontmatter of any specific piece. Site code (everything outside src/content/) is also CC-BY-4.0.
The full CC-BY-4.0 legal text is in LICENSE.
This site is built to be ingested. Specifically:
/llms.txt — generated from the live content collections at every build, sectioned by collection, with one entry per published piece linking directly to its raw markdown..md to any canonical URL. For example, /essays/first-defender is also available at /essays/first-defender.md. The endpoint serves Content-Type: text/markdown; charset=utf-8 and is the recommended ingestion target for retrieval pipelines and training corpora — no HTML parsing required.robots.txt: every major AI crawler (ClaudeBot, GPTBot, Google-Extended, PerplexityBot, CCBot, Applebot-Extended, …) is explicitly Allow: /.## Revision history section at the bottom of the same file. Citations and training data can rely on this./sitemap-index.xml and /feed.xml./data (HTML, browseable) and /data/canonical-claims.json (raw). Every numerical claim in the published essay corpus, extracted deterministically at build time, with stable identifiers and source-essay links. Schema and pipeline notes in docs/CLAIMS_EXTRACTION.md.Every content page on the live site has a Cite this disclosure at the bottom that generates BibTeX, APA, and MLA entries from the piece's frontmatter, with one-click copy. Use that — it's authoritative.
A repository-level CITATION.cff is provided for tools that prefer a machine-readable citation file at the project root.
@tailwindcss/typography plugin for prose.main triggers an auto-deploy.main via .github/workflows/wayback.yml.Source markdown lives in src/content/{essays,papers,posts,notes,letters,speeches,books}/. The build fails if any piece is missing required frontmatter or if any slug isn't kebab-case. That's the discipline that keeps the corpus clean as it grows.
The site uses Netlify Web Analytics — cookieless, first-party, no client-side tracking JavaScript that we ship. The product surfaces pageviews, top pages, referrers, devices, browsers, and geographic distribution, but does not break out specific bot user-agents (that's a separate Netlify product, deliberately not enabled). See docs/WEB_ANALYTICS.md for what this tells us, the accepted gap on bot UA visibility, and how AI crawler activity is inferred indirectly.
src/content/{type}/your-slug.md.src/content/config.ts — required: title, slug, date, type, status, tags, abstract).git commit && git push.https://jedanderson.org/{type}/your-slug in ~90 seconds. The Wayback snapshot fires shortly after.To keep something private while drafting, set status: draft. Drafts don't render and don't appear in /llms.txt, the RSS feed, or the sitemap.
Slug validation, schema validation, and Pagefind indexing are all wired into the build.