Cassiopaean Experiment Transcripts API

[KJS]

Hello everyone,

I’ve built a small project that serves Cassiopaean Experiment transcripts through an API (Application Programming Interface), using this forum as the data source. The tool:

- Parses the session‑index post to generate a complete list of sessions (so whenever a new session appears, it’ll be available)
- Retrieves individual session posts and strips out all XenForo formatting, returning clean Markdown or HTML transcripts

This is especially useful for AI analysis (e.g., docetl), with an MCP (Model Context Protocol) endpoint exposed for integration with Claude Code or Perplexity (although Perplexity truncates response from MCP tool calls). There's a very simple search capability that uses the Bleve engine for keyword search. Keep in mind, that this is currently working on a 2 vCPU 1GB RAM server, without any caching reverse-proxy, so be gentle as this might be fragile a little bit I originally made that for myself, but maybe someone will find it useful. If this is the case, I can upgrade the machine and add some neat features.

The website is Cassiopaea Experiment Transcripts

API specification: Cassiopaea Experiment Transcripts API Docs
Instructions for AI bots: https://cassiopaea.pub.org.pl/llms.txt
Session index in a content negotiated format: https://cassiopaea.pub.org.pl/transcripts
Session index in a markdown format: https://cassiopaea.pub.org.pl/transcripts?format=markdown
Latest session in a markdown format: https://cassiopaea.pub.org.pl/transcripts/2026-01-24?format=markdown
Example search query: Search results for "Trump" | Cassiopaea Experiment Transcripts

 

[palestine]

Hello I was thinking that with such data, at some point, you may have to tag the "Q:" and answers "A:", so as to have it in a sort of database system, because your website would have "grown". Could be a huge job (and I don't know if this is something already done).