Podcast transcription turns an episode's speech into readable text. First, look for a transcript in Apple Podcasts, Spotify or on the show's episode page. If you produce the podcast, use an existing transcript as your starting point or transcribe your own final audio, then check the words, speaker turns and timing before sharing it.
The useful output depends on your task: a listener may only need to find a passage, while a creator needs text that matches the published episode and works wherever it will appear.
Last checked: September 14, 2026.
Find the episode's existing transcript
On a supported iPhone, touch and hold the episode in Apple Podcasts and choose View Transcript. Apple also supports searching episode text and tapping a passage to play from there. Availability depends on language and catalog distribution; private RSS episodes outside Apple's catalog do not get transcripts. A new episode's transcript can arrive later than its audio. See Apple's transcript support page for current availability.
On Spotify, transcripts appear in the Now Playing view and on episode pages in the mobile app for shows that have them. Some are generated automatically, some are supplied by the creator, and not every show has one. See Spotify's transcript page.
If nothing appears in either app, open the show's own episode page and look for a transcript link. Search the show name, episode title and “transcript” together if the site is difficult to navigate. Check the episode number and guest before relying on a result; a similarly titled interview may be a different recording.
For a quotation, play the passage and a little of the conversation on either side. Keep the episode link and the timestamp with your note so you can return to the source. Finding readable text does not settle whether you may republish it: use the creator's stated permission or ask before distributing a full transcript of someone else's episode.
Creating a transcript from your own audio
Start with the final episode export, including the introduction, edits and closing segment listeners will hear. A transcript made from the raw interview can contain removed comments and timestamps that no longer match. If the final edit changes after transcription, review the affected passages again.
Choose a transcription workflow that accepts your file and produces editable text. If the destination needs timed captions, confirm its required format before choosing the tool; readable paragraphs alone will not provide timing.
For supported episode files on an Apple Silicon Mac, Paraspeech is one option for local speech transcription with a downloaded offline model. Optional rewriting is a separate processing step; choose Transcription Only when you want the recognized wording without AI rewriting. The audio-file transcription guide covers the file workflow, and the product documentation explains model and processing requirements. This route starts from your audio file, not a public episode link.
Try a passage containing a guest name, a speaker change and any difficult audio before processing the full episode. Compare the result with the recording to judge how much correction the workflow will need.
Review what a reader could misunderstand
Listen through the episode while reading the draft. Prioritize errors that change who said something or what they meant. A smooth sentence can still contain a wrong number, a missing “not,” or an answer assigned to the host instead of the guest.
Names and episode vocabulary. Build a short reference from the guest's confirmed spelling, organizations discussed and specialist terms. Use it to resolve uncertain words, then check replacements in context. A global replacement can fix a person's name while damaging an ordinary word elsewhere.
Speaker turns. Keep labels consistent and verify them against the voices. A label such as “Speaker 2” is not proof of identity. If two people overlap and you cannot establish the words, mark the unclear passage instead of assigning a plausible sentence to either person. Buzzsprout's transcript guidance also recommends speaker names, timestamps and relevant sound descriptions.
Meaningful sounds. Include a brief description when the sound explains the conversation: an alarm that interrupts an answer, for example. Avoid filling the page with descriptions of background noise that do not help the reader follow the episode.
Here is an illustrative editing example, not an actual transcription result:
Draft: “Maya: We can ship fourteen on Friday.”
After checking the audio: “Maya: We can't ship forty on Friday.”
Unresolved audio: “Maya: We can't ship [unclear quantity] on Friday.”
The correction changes both the commitment and the quantity. If the audio does not settle the quantity, leave it marked for review.
Keep a transcript faithful to the conversation. If you remove repetitions or false starts for readability, avoid rewriting a hesitant answer into a firm claim. Put an interpretive summary in a separate section or document.
Publish text that matches the episode
A transcript lets readers follow the conversation. Show notes can introduce the guest, link to resources and point to selected moments. A summary condenses the discussion. Give each a clear label so a reader looking for the full exchange does not mistake highlights for a complete transcript.
For a website, make the reviewed transcript easy to reach from the episode page. Include the episode title, participants and a link to the audio. Buzzsprout, for example, describes displaying transcripts on its hosted site and linking them through RSS; its guidance recommends linking from show notes rather than pasting the entire transcript into the episode description. See Buzzsprout's transcript workflow before uploading.
For Apple Podcasts, creators can provide transcripts through a supporting host's RSS feed; subscriber episodes accept VTT or SRT uploads in Podcasts Connect. Apple also lets creators download and correct transcripts. For a missing new transcript, its troubleshooting guidance says to allow at least 24 hours and inspect episode warnings. Follow Apple's submission and correction instructions.
For Spotify, creators with access can turn automatic transcripts on or off, download one, correct it locally and upload the corrected VTT or SRT file, up to 5 MB and with timestamps. An uploaded file is replaced if you later enable automatic transcripts. The feature is not available to every creator or in the Spotify for Creators mobile app, and depending on your podcast host you may not be able to sync transcripts. See Spotify's transcript management page.
Treat a timed file as a separate deliverable from the readable page. After an edit that removes audio, check timing after the cut as well as at the start. A correct opening timestamp cannot reveal a later offset. Spot-check the middle and end in the destination player, and inspect the text there for missing paragraphs or broken speaker labels.
Finally, open the episode as a listener would. Follow its transcript link, find a passage and compare it with the audio. If the text refers to a different edit, correct that mismatch before sharing the link.
