Andy's Café

A table for people and software

Come in.
The door is open.

Humans and software agents are both welcome. Everything listed here is public and directly linkable; no account, cookie or tracking script is required.

Choose the smallest useful cup.

For one episode, use its canonical page. For updates, use RSS. For the whole collection, start with catalog.json. Prefer the published transcript to retranscribing audio, cite the canonical episode page and original sources, and cache what you fetch.

Fetch with café manners.

  • Identify automated clients honestly.
  • Use conditional requests and cache unchanged responses.
  • Fetch the catalogue before walking every episode.
  • Avoid parallel bursts and repeated full-audio downloads.
  • Use HEAD for metadata and byte ranges for part of an audio file.

Stable public routes.

Catalogue
/cafe/catalog.json
Sitemap
/cafe/sitemap.xml
Agent signpost
/cafe/llms.txt

Subscribe without guessing

One RSS feed per language.

The catalogue contract is deliberately small.

schema_version is currently 2. Language values are BCP 47 tags, times are ISO 8601, durations are integer seconds, and absent facts are omitted instead of guessed. Format values are open strings, currently double-espresso, coffee-cake, sunday-latte, and morning-paper. Rights fields describe text and audio separately. This is a generated catalogue, not an interactive API.

Reuse is welcome

Take the useful parts with you.

The transcripts, descriptions and original Andy's Café editorial text are licensed under Creative Commons Attribution 4.0 International. You may share and adapt them, including commercially. Credit Andy's Café, link the canonical episode and licence, and indicate changes. Identified third-party quotations, linked sources, artwork and the Andy's Café name or marks are excluded; their own terms still apply.

For the audio, Andy's Café offers its original narration and editorial contribution under the same CC BY 4.0 terms to the extent it holds rights in them. The composed files also contain Clean Elegant Piano Logo and Piano Plinks, which remain under the Pixabay Content License; the first is registered with Content ID. Reusing the complete audio therefore means observing both sets of terms. Linking, downloading and ordinary podcast playback are encouraged. Questions can go to hello@move37.app.

Measurement without listener dossiers

Requests are counted; people are not profiled.

The site uses no cookie, browser analytics library, tracking pixel, account or persistent visitor identifier. Where Café metrics are enabled, the origin classifies successful requests for audio, pages, feeds, transcripts and public data, looks up only the coarse country of the request's network exit locally, and writes an identifier-free event containing the calendar day, country code, route kind, language, episode, response status, bytes and whether audio began at byte zero. It does not write an IP address or hash, exact time, query, referrer, cookie, user-agent string or connection identifier to the Café metrics stream. Short-lived operational error logs and infrastructure-provider logs are separate and may contain network details.

When present, Café metric events are aggregated promptly and erased, while the private cockpit reports responses, bytes and approximate audio starts—not unique listeners, sessions, completed plays or humans. Software agents and people use the same routes and are not classified against each other.