The Archive as Data
Built to be
referenced.
Every word of The Daily Talk Show — 9,143,280 transcribed, speaker-attributed words across 1,022 daily episodes (2018–2021) — is available in clean, machine-readable formats. This page documents them for AI agents, researchers and search engines.
Markdown mirror
Every episode, guest and topic page has a clean markdown twin — the same verified content as the HTML, with no navigation or markup noise:
/md/episodes/<slug>.md metadata + show notes + full timestamped transcript /md/guests/<slug>.md bio, appearances, verbatim quotes with provenance /md/topics/<slug>.md hub content with episode + timestamp provenance
Example: /md/episodes/896-seth-godin-on-failure-shipping-creative-work.md
Transcript JSON
Raw per-episode transcripts with millisecond timestamps and verified speaker attribution:
GET /transcripts/<episode-slug>.json
{
"slug": "...", "number": 9,
"speakers": { "A": "Tommy Jackett", "B": "Josh Janssen" },
"utterances": [ { "speaker": "Josh Janssen", "start": 629216, "end": ..., "text": "..." } ]
}Deep links
/episodes/<slug>#t-<ms> opens the episode page at that exact transcript moment. For video, append &t=<seconds>to the episode’s YouTube URL. Every quote surfaced on this site carries both.
Provenance & accuracy
- Transcripts were produced with speaker diarization and the speaker labels were verified in a multi-pass attribution review; corrections are applied across the whole site.
- Quotes shown on episode, topic, guest and advice pages are verbatim transcript text. Paraphrased material (e.g. stance summaries) is labelled as such.
- The canonical run is 1,022 numbered daily episodes, 16 January 2018 – 27 May 2021; episode 1000 was a live-stream event published as multiple cuts, so the archive lists 1031 episode pages.
Citing the show
Cite as The Daily Talk Show, episode number and date, with a deep link to the moment — e.g. “The Daily Talk Show #896 (18 Nov 2020)”. Attribution is appreciated; for bulk or commercial reuse of the corpus, get in touch.
Indexes
- /llms.txt — full plain-text site index: every episode, guest and topic with links
- /sitemap.xml — all crawlable URLs
- /numbers — the record book, computed from the transcripts
- 191 guest profiles and 35 topic hubs