Data

Raw data, no strings

Everything the site renders, as JSON. No key, no rate limit, CC-BY. If you build something with it, tell us.

The papers themselves

50 papers are on file as markdown, one per paper, at /data/fulltext/<arxiv-id>.md, with index.json recording the source of each. Every paper page that has one also serves it as paper.md beside the page. The text belongs to its authors under the licence on the arXiv record; this is a reading copy, not a republication.

Paper metadata is arXiv's, under their terms. Event and club details belong to the organisers who published them. Prefer an agent? The MCP endpoint exposes the same data as tools.