Skip to main content

Find what someone said

GET /quotes answers quotes from podcasts, long interviews, hearings and companies’ own events, each with its exact words, its speaker, the show and episode, the day, the second it starts, and a page to cite.
  • q matches by meaning, best match first. Words in quotation marks match exactly, in order.
  • Without q, the newest episode comes first.
  • Narrow it with person (an id from GET /people), company (the speaker’s), named (a company or person the quote names), ticker, show, kind (CEOs, investors and the rest), from and to, and list, one of your team’s lists.
  • A quote whose speaker isn’t confirmed is left out unless confirmed=false, and then it’s marked so.
  • format=csv gives the list as a file.
Each quote has text, its words; speaker; show and episode; starts_at, the second; url, the page to cite on poddata.ai; and cite, a line to quote it by. GET /quotes/{id} gives one, and GET /quotes/{id}/source the recording from its second, with who spoke when. Cite a quote by its url and its cite line, with the speaker and the day. A key gets at most 20 different quotes from one episode a UTC day: see Keys and limits.

Follow a topic or a person

A tracker follows what a sentence describes, such as “CEOs on power limits for data centers”, and sends each new quote it matches.
  1. Preview it. See what the sentence would have caught in the last 90 days, before making it:
    A team has 60 previews a UTC day.
  2. Show the person what it caught, and change the sentence until they’re happy with it.
  3. Make it, with a name:
  4. Read what it matches with GET /trackers/{id}/matches, or every tracker’s together with GET /matches. Mark a wrong one with POST /matches/{id}/feedback, and the tracker learns from it.
A match reaches the team by Slack, Microsoft Teams or a webhook. A tracker’s email is by request today. A plan sets how many trackers a team may have running: see Keys and limits. An assistant asks the person before it makes, changes or deletes a tracker.