Guide · Podcasts

How do podcasts get cited in AI answers?

Updated 2026-11-10 · kagentx

Short answer

Make a web page for every episode on your own site, with show notes and the full transcript, and add PodcastEpisode structured data. Tested on 2026-11-10: Spotify's show page has the description, rating and recent episode descriptions but no transcript text; Spotify's robots.txt blocks training crawlers across the site while leaving /embed/ open to OAI-SearchBot, ChatGPT-User, Claude-User, PerplexityBot and others; we could not fetch an Apple Podcasts show page, so that is not verified. Whether AI reads platform transcripts is not stated, and we have no evidence.

What can AI read on a podcast platform's show page?

Text, but not what is said in the audio. On 2026-11-10 we tested the robots.txt of Apple Podcasts and Spotify, one Spotify show page and one public RSS feed: the Spotify show page's HTML has the show description, the rating and the descriptions of recent episodes, but no transcript text; we could not fetch an Apple Podcasts show page, so that is not verified. Whether AI cites podcast platform pages we have no evidence, so it is not verified.

Platformrobots.txt and page (2026-11-10)What AI can read
Apple Podcasts (podcasts.apple.com)robots.txt blocks only /WebObjects/, /api/, /includes/, /v1/ and names no AI crawler; lists 5 sitemaps for shows, episodes, channels and more; the show page was redirected from our network and could not be fetchedWhether the show page's HTML has the description, the episode list and a transcript: not verified
Spotify (open.spotify.com)robots.txt blocks the whole site for training crawlers such as GPTBot, ClaudeBot, anthropic-ai, CCBot, Google-Extended and Applebot-Extended; OAI-SearchBot, ChatGPT-User, Claude-User, PerplexityBot, Perplexity-User and others have their own groups containing only Allow: /embed/ and Allow: /oembedThe show page HTML (about 358 KB) has the show name, publisher, description, rating, categories, and the titles, descriptions and durations of recent episodes; the JSON-LD is PodcastSeries with no episode list; no transcript text
Your RSS feedA public XML file; the feed we tested has 3,032 episodes and is 20.6 MBChannel and per-episode title, description, show notes (content:encoded), date, duration; the feed we tested has no podcast:transcript

Spotify's robots.txt is worth a closer look: it blocks training crawlers across the site but leaves /embed/ and /oembed open to search and user-triggered crawlers, and a comment in the file says this is so the embedded player can be discovered. You can draw the same line on your own site; the rules are in How to write robots.txt for AI crawlers.

What do Apple and Spotify officially say about transcripts?

Both generate transcripts for shows, but both are for listeners inside the app. Below is what the docs say (viewed 2026-11-10). The docs do not say whether these transcripts are public to search or AI crawlers, and we did not find them in the HTML of a Spotify show page, so it is not verified.

PlatformDocsWhat it means for you
Apple PodcastsA transcript is generated automatically after a new episode is published; the episode detail page has a plain-text transcript; listeners can copy up to 200 words when sharing; your own transcript is supplied through the RSS transcript tag, as VTT or SRT; iOS 17.4 and later, 11 languages supportedYou can point to your own transcript file in RSS so the proofread version is shown (with speaker names)
SpotifySome shows have auto-generated transcripts, shown on the player and episode pages in the mobile app; you can upload VTT or SRT, up to 5MB; turning on auto-generation replaces what you uploaded; putting the transcript in RSS makes it available elsewhere, but some platforms do not support itThe transcript lives in the app; to make it readable on the web, host it on your own site

Conclusion: platform transcripts are a listener feature, and no official statement says whether AI can read them. The only text we can be sure crawlers read is in two places: the description and show notes on the platform page, and the pages on your own website.

Why should your own site host a transcript page and show notes for every episode?

Because when AI answers “who said what on podcast X” or “recommend a podcast about Y”, the only thing it can cite is text. The audio itself is not read, and the episode description on a platform page is often only a few lines. One page per episode on your own site, with show notes and a transcript, is the only place you fully control and crawlers can read directly.

PageQuestion it answersWhat goes in
Show home pageWhat is this podcast and who is it for?Define it in the first paragraph: topic, hosts, release frequency, language; the latest episodes; links to each platform and the RSS address
Episode pageWhat is this episode about, and who said what?Title, date, duration, guests; a three-to-five-sentence summary; chapters and timestamps; related links; an audio player; the full transcript below
TranscriptWho said a given line, and when?Split by speaker with timestamps; use plain HTML text, not a PDF or an image, and not loaded by script
Guest pageWhich episodes has this guest been on?Guest bio, their website, links to the episodes they appeared in

This is like video: the video itself is not read, and the title, description and transcript are the material; see Are YouTube videos cited in AI answers?. The same goes for newsletters, where the article page is what gets cited; see How do I get my newsletter cited by AI? Substack, beehiiv and Ghost compared.

How do you write PodcastEpisode and PodcastSeries structured data?

schema.org (version 30.1, viewed 2026-09-16) has the types PodcastSeries and PodcastEpisode. An easy mistake: the transcript property is used only on AudioObject and VideoObject, so put the transcript inside the episode's associatedMedia (an AudioObject) rather than directly on PodcastEpisode. PodcastSeries has a webFeed property whose value is the RSS URL. Structured data is only a supplement and must match the text on the page; see Does structured data help AI search? for what it does.

{
  "@context": "https://schema.org",
  "@type": "PodcastEpisode",
  "name": "Episode 42: How we cut our build time in half",
  "url": "https://example.com/podcast/42/",
  "datePublished": "2026-11-03",
  "episodeNumber": 42,
  "duration": "PT38M",
  "description": "One sentence that matches the first paragraph of the show notes.",
  "partOfSeries": {
    "@type": "PodcastSeries",
    "name": "Example Show",
    "url": "https://example.com/podcast/",
    "webFeed": "https://example.com/podcast/feed.xml"
  },
  "associatedMedia": {
    "@type": "AudioObject",
    "contentUrl": "https://example.com/audio/42.mp3",
    "transcript": "The full transcript text goes here, the same text that is on the page."
  }
}

Names and URLs in the example are placeholders. Whether Google or any AI product uses this markup is not stated, and we have no evidence.

How do you make a podcast easier for AI to cite, step by step?

  1. 1. First confirm AI crawlers can reach your site

    Use AI crawler check to see whether robots.txt blocks search crawlers.

  2. 2. Make a web page of your own for every episode

    Title, date, guests, summary, chapters and related links, written directly in the HTML.

  3. 3. Put the full transcript on the episode page

    Split by speaker, with timestamps, and proofread names, product names and numbers.

  4. 4. Point to the transcript in RSS and write good show notes

    Apple Podcasts reads your transcript (VTT or SRT) through the RSS transcript tag; put show notes in content:encoded, with the first sentence saying what the episode is about.

  5. 5. Add PodcastSeries and PodcastEpisode markup

    Fields match the page, and the transcript sits in the AudioObject inside associatedMedia; see the example above.

  6. 6. Test again in two months with the same questions

    Ask ChatGPT and Perplexity “what podcasts cover X”, and also ask about specifics from one or two episodes; note whether your page is cited; if not, go back to step 3.

FAQ

Will ChatGPT cite Spotify and Apple Podcasts pages?

We found no official statement and have no evidence of actual citations, so it is not verified. What we can confirm: Spotify's robots.txt has no disallow rule for OAI-SearchBot and ChatGPT-User, only Allow: /embed/ and /oembed; Apple Podcasts' robots.txt blocks only /WebObjects/, /api/, /includes/, /v1/ (2026-11-10).

Can AI read the transcripts the platforms generate automatically?

Not verified. The docs from Apple and Spotify (viewed 2026-11-10) both present transcripts as a feature for listeners in the app and do not say whether they are public to crawlers. We did not find transcript text in the HTML of a Spotify show page.

How do you give Apple Podcasts your own transcript?

The docs (viewed 2026-11-10): supply it through the RSS transcript tag as VTT or SRT; if you do not, Apple generates one automatically after a new episode is published.

Can the transcript property go on PodcastEpisode?

In schema.org, transcript is used only on AudioObject and VideoObject, so we suggest putting it in the episode's associatedMedia (an AudioObject), with text that matches the page.

Which RSS fields are public text?

In one public feed we tested (2026-11-10), the channel has title, description, language, author and category; each episode has title, description, show notes with HTML (content:encoded), date, duration, itunes:subtitle and itunes:summary.

Sources

kagentx

Can crawlers read your pages?
Let the evaluation tell you

The evaluation checks whether crawlers can get in and whether the text is served directly, then asks ChatGPT, Perplexity and Claude (with web search on) your buyers' questions and sees whether you are mentioned. The Basic plan is $1, and results usually take 2–3 minutes.