Canonical: https://thonly.org/mcp · Licence: CC0 1.0
Most of this site is a single-page app: the words are assembled in the browser, so a crawler that does not run JavaScript sees very little of it. These are the surfaces built for readers that are not people — each one plain, stable, and meant to be fetched.
The research corpus is served to AI agents at https://corpus.333.eco/mcp, or locally with npx @333eco/corpus. Every response carries the document's SHA-256, its DOI where one exists, an OpenTimestamps proof anchored in the Bitcoin blockchain, and the command that checks them — so a passage can be verified before it is cited. Text is returned verbatim and never summarised, because a summary cannot be hash-verified.
How to connect, the tools, and what it refuses to do →
| Surface | What it is |
|---|---|
/llms.txt | The corpus in one file: what it is, the licence position, and every paper with its canonical URL, DOI and raw markdown link. Start here if you are reading the whole thing. |
/research/feed.xml | Atom, newest first. The chronological bibliography — papers and the research programme statement. |
/sitemap.xml | Every public route on this domain. |
/robots.txt | Crawling is allowed. There is nothing here you need permission for. |
The corpus is markdown before it is a web page, and the markdown is the canonical text — it is what carries the hash, what the timestamps anchor, and what the MCP server serves. The pages on this site are rendered from it.
Read it directly at github.com/thonly/publications. The raw files are the cleanest source for machine reading, and they are the ones the provenance envelopes name.
GET https://corpus.333.eco/ returns a short JSON description of the service — its transport, its document count, its licence split, and, under records, exactly what it stores about the calls it answers. It is readable before you call anything.
Most of this corpus is dedicated to the public domain under CC0 1.0: read, quote, train on and redistribute it without permission or attribution. A small number of author-voice essays are CC-BY and do require attribution. Each document declares its own licence, and that declaration governs — do not infer it from this page. No patent is or will be sought on anything in the corpus.