Paste an episode link and JotVox transcribes the audio into a timestamped script you can quote, search, and turn into show notes.
Get a Apple Podcasts transcript →Not as a file you can take somewhere else. Apple's own app shows a transcript view for some episodes, but that is a reading feature inside the app. It is not a download, not an SRT, and not something you can drop into a doc, a CMS or a research folder.
So for most people the practical route is to transcribe the audio. JotVox takes the episode link, runs the audio through AI transcription and hands back text you own. There is no caption layer to grab first the way there is on a video platform, so every podcast episode uses a credit.
What you get in exchange is more than a reading view: timestamps throughout, real export formats, and the option to run an AI summary on top of the transcript.
This is what most people are here for, and it is worth spelling out as a sequence rather than a promise.
A raw two hour transcript is not show notes. It is the raw material for them. The order that works:
Each AI action is one credit, so a full workup on an episode usually costs two: one to transcribe, one to summarise. There is a longer walkthrough in building show notes from a transcript.
Podcasts are long. One to three hours is normal, and some interview shows run past that. Length alone changes what a transcript is for. Nobody reads 25,000 words end to end, so chapters and search matter far more here than they would on a ten minute clip.
Then there is the room. Most shows have two or more voices, and real conversation involves interruption, crosstalk and someone laughing over the answer. A remote guest on a laptop mic sounds nothing like the host on studio gear in the same file. The model handles this well, but a moment where three people speak at once is genuinely ambiguous in the audio, and the transcript will show that.
The recurring problem is names. Guest names, company names, book titles and technical jargon are exactly the words you most want right and the ones a model is most likely to render phonetically. Check proper nouns before you publish a quote. The rest you can usually trust.
Ad reads and sponsor breaks come through as spoken text like everything else, because that is what they are in the audio. If you are cutting show notes, they are the first thing to delete.
A timestamped transcript changes what an episode can be used for. You can quote a guest word for word instead of paraphrasing from memory, and send a reader to the exact minute. Journalists and researchers get a source they can cite. Anyone building an internal knowledge base gets an archive a search box can reach into, which an audio file will never be.
Exports handle the handoff: TXT for writing and research, SRT or VTT when the episode is being cut into video clips, and MP3 when you want the audio itself. If you publish those clips short-form, the TikTok and Instagram pages cover pulling text back off them afterwards.
Subscriber-only and paid episodes. If an episode sits behind a paid podcast subscription, its audio is not publicly reachable and a link will not work. If you subscribe and your app lets you download the file, upload that file to JotVox instead.
Region-limited shows. Some catalogues are not distributed everywhere. Audio that cannot be reached cannot be transcribed.
Pulled episodes. Shows delete episodes. A link in an old newsletter may no longer resolve to anything at all.
Very long episodes. These do work, and one credit still covers one transcription whatever the runtime. But a four hour transcript is only useful with a summary layered on top, so budget the second credit. If the show also mirrors episodes to SoundCloud, that link works the same way.
1. Open the episode and copy its link Find the episode in Apple Podcasts on your phone or at podcasts.apple.com. Use the share option on the episode itself rather than the show, and copy the link. An episode link points at one recording; a show link points at the whole feed.
2. Paste the link into JotVox Drop it into the box at jotvox.app. If the episode is paid or subscriber-only, download the audio where your app allows it and upload that file instead, from your device, Google Drive or Dropbox.
3. Transcribe the audio Podcasts have no caption file to lift, so JotVox runs AI transcription across the whole episode. Language detection covers 99 languages. This uses one credit.
4. Summarise, then export Run an AI summary for chapters, takeaways and quotes, which is what makes a two hour episode usable. Then export as TXT, SRT, VTT or MP3.
No. Apple's app has a transcript reading view for some episodes, but it stays inside the app. You cannot export it as a file, paste it into a document, or hand it to an editor. To get text you own and can reuse, transcribe the episode audio. JotVox returns a timestamped transcript you can download as TXT, SRT or VTT.
One credit for the transcription, whatever the runtime. A 20 minute episode and a three hour episode each count as one action. An AI summary with chapters and takeaways is a second action, so a complete show notes workup usually costs two credits. New verified accounts get three free credits plus one free AI transcription per day, and credit packs start at $4.99 for 20.
Treat speaker attribution as something to check rather than assume. On a clean two person interview with separate microphones, the transcript usually tracks the conversation clearly. On a panel show with crosstalk, overlapping speech is genuinely ambiguous in the recording itself. Read the first few minutes to orient yourself, then correct any attribution before you publish a quote from it.
A link on its own will not work, because the audio is not publicly reachable. If you subscribe to the show and your podcast app lets you download the episode as a file, upload that file to JotVox from your device, Google Drive or Dropbox. The transcription then runs exactly the same way. Without access to the audio in some form, no tool can transcribe it.
Yes. AI transcription covers 99 languages and JotVox can detect the spoken language rather than making you choose. That handles most non-English shows, including bilingual episodes where a host switches between segments. Heavy code-switching inside a single sentence is harder than a clean change between speakers. Accuracy is highest on clear studio audio in any of the supported languages.