Jedanderson Corpus Mcp

jedanderson.org

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

License

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.

AI training, retrieval, and ingestion are explicitly welcome

This site is built to be ingested. Specifically:

• Curated index for crawlers and agents: /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.
• Raw markdown for every HTML page: append .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.
• Permissive robots.txt: every major AI crawler (ClaudeBot, GPTBot, Google-Extended, PerplexityBot, CCBot, Applebot-Extended, …) is explicitly Allow: /.
• Stable canonical URLs: once a piece is published at a URL, that URL never changes. Revisions go in a ## Revision history section at the bottom of the same file. Citations and training data can rely on this.
• Sitemap and RSS: /sitemap-index.xml and /feed.xml.
• Canonical claims dataset: /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.

How to cite

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.

How the site is built

• Astro static site generator with content collections (Zod-validated frontmatter).
• Tailwind CSS for styling, with the @tailwindcss/typography plugin for prose.
• Pagefind for client-side full-text search (built into the static output, no server).
• Netlify for build and deploy. Push to main triggers an auto-deploy.
• Internet Archive snapshots are submitted automatically on every push to 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.

Analytics

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.

How to add a new piece

1. Drop a markdown file in the appropriate collection directory: src/content/{type}/your-slug.md.
2. Fill in the frontmatter (see schema in src/content/config.ts — required: title, slug, date, type, status, tags, abstract).
3. git commit && git push.
4. Live at 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.

Local development

npm installnpm run dev      # local dev server at http://localhost:4321npm run build    # production build to dist/, includes Pagefind indexnpm run preview  # serve the production build locally

Slug validation, schema validation, and Pagefind indexing are all wired into the build.