Preface
You recorded a good episode. Fifty-eight minutes, one guest, and three ideas in it that you have not heard anyone else say out loud. It published on Tuesday. By Friday it was six items down the feed, and the only written record of it is a two-line description you typed in a hurry.
The write-up would have carried that episode much further. A real set of show notes. A blog post built from the best fifteen minutes. Four pull quotes for social. Captions for the video cut. None of it happened, and the reason is not that you were lazy.
It did not happen because all of it depends on one thing you did not have: a transcript. Without a transcript you are scrubbing back and forth through an hour of audio trying to find the bit where your guest said the thing about pricing. With a transcript, the whole job becomes editing, which is a job you can actually finish.
Why the write-up never happens
Ask ten podcasters why the show notes are thin and you get the same two answers.
The first is money. Transcription is usually sold by the minute or by the hour. A weekly show at an hour an episode is roughly fifty hours of audio a year, and every one of those hours is metered. The meter changes how you behave long before it empties your wallet. You start deciding which episodes are worth transcribing, and the answer becomes "the big ones", which means most of your back catalogue stays as sound and nothing else.
The second is trust. To use most transcription services you upload the file. That is a decision you make about an episode that has not gone out yet, sometimes with a guest who spoke freely because they believed there were two people in the room. Off-the-record asides, an unannounced product, a name they mentioned and then asked you to cut: all of it goes up with the file. Most of the time nothing bad happens. That is not the same as the material staying yours.
Put the two together and the pattern is predictable. The transcript does not get made, so the article does not get written, so the episode is heard by the people who already subscribe and by nobody else. Meanwhile the words inside it are the most searchable thing you produced all week.
Without a transcript
The episode dies in the feed
One asset, one format, one audience. The good material is locked inside an audio file that no search engine can read and that you cannot skim. Two weeks later you remember there was a great line about hiring in episode 41 and you have no way to find it except listening again.
With a transcript on your machine
One recording becomes four assets
Show notes, a written article, a handful of pull quotes, and captions for the video cut. All four come out of the same text, and the text is searchable, so episode 41 is one query away for the rest of your life.
What this workflow is, and what it is not
The workflow below uses Yaps Studio to transcribe the episode on your own computer, and Yaps dictation to draft the write-up by talking instead of typing. Both run on your machine. Nothing about the recording is uploaded, which means an unreleased episode stays unreleased.
Be clear about the boundaries before you start, because they shape the whole process.
Yaps does not record your podcast. It is not a remote recording studio, it does not join a call, and it does not sit in on your interview. You keep your own setup, whatever that is, and you bring the finished audio or video file to Yaps afterwards. Studio works on a file you already have.
Yaps also does not publish anything. There is no integration with a podcast host, a content management system, or a social scheduler. What you get out of it is clean, paste-ready text and exported files: plain text, Markdown, WAV audio, and SRT subtitles. Pasting those into your host or your blog is a manual step, and it takes about a minute.
And Yaps does not write the article for you out of nothing. It gives you an accurate record of what was actually said, and then it lets you draft over that record in your own voice at speaking speed. The thinking is still yours. That is the point: an article assembled from your real words sounds like you, because it is you.
This guide is only about the write-up. If you are choosing voice tools for the whole production pipeline, including scripting and voiceover, our guide to voice tools for podcast production covers the earlier stages.
How to turn an episode into an article in six steps
The first pass takes an evening. Every episode after that is faster, because you stop re-deciding how to do it.
1. Keep the recording as a file
Record however you already record. A local multitrack session, a remote recording tool, a phone in the room, a video call you exported: it does not matter, as long as you end up with an audio or video file sitting on your computer.
This is worth stating plainly because it is the most common misunderstanding. Yaps does not record your podcast, does not join a call, and does not act as a remote studio. It works on the file after the fact. If your recording tool gives you separate tracks per speaker, keep them, but a single mixed file works fine.
2. Switch on speaker identification for a multi-voice episode
If your show is one host talking, skip this step. If it is two hosts, an interview, or a panel, a transcript that does not say who spoke is much less useful, because every quote you pull needs attribution.
Yaps can label the turns for you, and there are three precise conditions. It is a desktop feature on Windows and macOS only, so it is not available on Android or Linux. It is optional and off by default. And it needs a one-time download: open the Features screen in the app, install the speaker identification pack, which is around 32 MB, then switch on "Identify speakers" before you start the transcription. Turning it on afterwards does not retroactively label a transcript you have already made.
The labelling runs on your device like the rest of the job, so a three-voice interview still never leaves your machine. Expect to correct a label here and there when you read through. That is a fraction of the work of marking every turn by hand.
3. Import the episode into Studio and transcribe it offline
Open Studio, import the episode file, and start the transcription. There is no upload step, so the work begins immediately rather than after a progress bar finishes sending an hour of audio to somewhere else.
Two things follow from the work happening locally. The first is privacy: an unreleased episode, an interview with an embargo on it, or a guest conversation you have not yet cut stays entirely on your disk. The second is cost shape. Because this is a flat subscription rather than a meter, a fifteen-minute episode and a ninety-minute episode cost the same, which is to say nothing extra. That is what makes it reasonable to go back and transcribe the whole archive on a quiet afternoon.
If you want the mechanics of file transcription in more detail, including formats and what affects quality, we cover them in the guides to turning a video into a transcript offline and transcribing audio to text.
4. Read the transcript once and fix the names
Do not skip this. Read the transcript through once at speed, with the audio available for the parts that look wrong.
You are looking for three things. Proper nouns, because names of people, companies, and products are where any transcription goes astray. Technical terms specific to your show. And misplaced speaker labels, where a short interjection got attached to the wrong voice. Fifteen minutes of cleanup protects you from publishing a quote your guest never quite said, which is the one mistake in this whole workflow that actually costs you something.
5. Draft the article by talking over the transcript
Here is the move that changes how long this takes. Do not type the article, and do not try to edit the transcript into an article. Put the transcript on one side of the screen and a blank note on the other, then press the Yaps hotkey and talk.
Explain the episode out loud as if a friend asked what it was about. Glance at the transcript as you go and lift the lines that carry the point. You will find yourself saying "and this is the bit that matters" and then reading the quote, which is exactly the shape a good write-up has anyway. On desktop the hotkey is Fn, and it works in whatever app your note happens to be in.
Dictated drafts read better than typed drafts for this particular job, because a podcast write-up is meant to sound spoken. Text cleanup runs on your device as you talk: it strips the filler words and false starts, fixes punctuation and capitalisation, and formats lists, so what lands in the note is a draft rather than a raw stream. You still edit it. You are editing something that already exists, which is a much smaller task than facing an empty page.
6. Export the pieces and paste them where they go
Keep the transcript and the drafts in Yaps notes, where local search finds them again. Notes handle plain text, checklists, and kanban boards, so a per-episode note with a checklist of assets to publish is a natural fit.
Then export. Notes come out as Markdown or plain text, which paste cleanly into a podcast host, a blog editor, or a newsletter. Studio exports SRT subtitles for the video version of the episode, and WAV audio if you need it. Those are the formats: WAV and SRT, not MP3 and not VTT.
The pasting is manual, and it is the last honest limitation of this workflow. Yaps has no integration with any podcast host, content management system, or social scheduler. It produces finished text and files, and you place them.
What this costs you in time
An hour-long episode does not take an hour to write up. Here is the realistic shape of a session, once you have done it twice.
Set up speaker labels2 min, once
Install the speaker identification pack from the Features screen on Windows or macOS and switch on "Identify speakers". You only do this the first time.
Transcribe the episodeHands off
Import the file into Studio and start it. No upload, so it begins straight away and you can go and do something else while it runs.
Read through and fix names15 min
One skim for proper nouns, product names, and any speaker label sitting on the wrong turn. This is the step that keeps your quotes honest.
Dictate the show notes5 min
Five or six spoken sentences on what the episode was about, cleaned up on-device as you talk.
Dictate the article draft20 min
Talk through the episode with the transcript open beside you, lifting the lines that carry the argument. A thousand words arrives faster than you expect.
Edit, export, paste25 min
Tighten the draft by hand, pull four quotes out of the transcript, export SRT for the video cut, and paste everything where it belongs.
Add those up and it is a little over an hour of your attention for four assets, and only twenty-five minutes of it is drafting, because the two dictation passes are the fast part. Compare it to the version where you type from playback, pausing every fifteen seconds to catch a sentence, and the difference is most of an afternoon.
The four things one episode becomes
The transcript is not the deliverable. It is the raw material for four separate things, and it is worth knowing the shape of each before you start talking.
Show notes. Five or six sentences on what the episode covered, a list of the topics in order, and any links or names mentioned. The list of topics is the part people actually use, because it tells a browsing listener whether this episode is for them. Dictate this first, while the episode is still in your head.
The article. Not a summary of the episode, and definitely not the transcript with the "ums" taken out. Pick the single strongest idea in the hour, argue it in your own words, and quote your guest where their phrasing beats yours. Eight hundred to fifteen hundred words is plenty. This is the asset that brings in people who have never heard of your show, because it is the only one a search engine can read.
Pull quotes. Four short lines, verbatim, with attribution. These are for social posts and for the episode page. Having the transcript is what makes them accurate, and accuracy matters when you are putting words in a guest's mouth in public.
Captions. If you publish a video version, export SRT from Studio and upload it alongside the video. Captions make the episode watchable with the sound off, which is how a large share of people watch anything on a phone.
An episode with no transcript is an hour of your best thinking that nothing can search, skim, or quote.
Yaps
How the options compare
There are three realistic ways to get from an episode to a written write-up. Here is the honest split, including where the cloud service still has the edge.
| What matters | Yaps (on-device) | Per-minute cloud service | Typing from playback |
|---|---|---|---|
| Where the episode goes | Stays on your disk | Uploaded to a server | Stays on your disk |
| Safe for an unreleased episode | Yes | Your call to make | Yes |
| Cost of a long episode | Same as a short one | Billed by the minute | Hours of your time |
| Works with no internet | Yes | No | Yes |
| Speaker labels | Yes, Windows and macOS | Yes | You type them |
| Subtitle export | SRT | SRT and other formats | No |
| Draft the article by voice | Yes, in any app | Transcription only | No |
| Searchable archive of every episode | Yes, local search | In their web app | Only what you typed |
| Publishes to your host | No, paste-ready text | Some integrate | No |
Two rows deserve a second look.
The cost row is the one that changes your behaviour. A meter makes you triage: you transcribe the interview with the big guest and skip the other three. A flat subscription running on hardware you already own removes that calculation, so you transcribe everything, including the archive, and search it later. That archive is quietly the most valuable thing this workflow produces.
The publishing row is where the cloud services win. Some of them push a transcript straight into your host or your video platform. Yaps does not, and will not pretend otherwise. If a one-click handoff to your publishing stack is the thing you care about most, that is a real reason to choose differently.
Where Yaps is not the right tool
Four cases, stated plainly.
If you need something to record the conversation for you, Yaps is the wrong tool. It does not record your podcast, does not join calls, and is not a remote recording studio. Keep whatever you use for that.
If you want live transcription while the conversation happens, or a bot that sits in a call and produces notes afterwards, Yaps does not do it. Studio works on a file you already have, after the fact.
If publishing integrations are the deciding factor, a service that connects directly to your podcast host or CMS will save you the paste. Yaps gives you clean text and exported files and leaves the placing to you.
And if you work from an iPhone or iPad, the desktop workflow is not available to you yet. An iOS version of Yaps is coming, and it is not here. Android is out today and free with no account, and it handles voice typing, voice notes, and read-aloud, but Studio file transcription and speaker identification are desktop features.
Who this is for
Solo podcasters get the most out of it, because they are the ones doing every job. The write-up is the task that slips when you are also editing audio, chasing guests, and cutting promos, and this is the version of the write-up that fits in an evening.
Interview shows get the second biggest win, from speaker labels. An accurately attributed transcript of a good conversation is a document with real value, and it is the difference between quoting a guest properly and paraphrasing them and hoping.
Anyone sitting on an archive of recordings should try this on the back catalogue before the next episode. Fifty transcribed episodes turn into a searchable body of your own thinking, which changes how you plan future shows. You stop wondering whether you already covered something and start checking.
And anyone whose recordings are sensitive before release, including journalists working with sources and researchers with interview material, gets the part that no per-minute service can offer: the file never leaves the machine. If that describes you, the dictation feature page and the transcription guides linked above go deeper on how the on-device side works.
The short version
Transcribe the episode in Studio on your own computer, turn on speaker identification first if there is more than one voice, read the transcript once to fix the names, then dictate the show notes and the article over it instead of typing them. Export Markdown for the text and SRT for the video cut, and paste them where they go.
Start with Yaps if the recording is yours and you would rather not hand an unreleased episode to a third party to get its text, and if you publish often enough that a per-minute meter would make you skip episodes. Reach for a cloud service instead in the specific case where a direct integration with your podcast host matters more than keeping the audio local, or where you need a bot to sit in a live call. For the ordinary weekly job of turning a recording you already have into words people can find, an on-device transcript and a dictated draft is the right default.
Frequently Asked Questions
How do I turn a podcast episode into a blog post?
Transcribe the episode first, then draft the post over the transcript rather than from memory. With Yaps you import the recording into Studio, transcribe it on your own computer, read through once to fix names, then press the Yaps hotkey and dictate the article while glancing at the transcript for the lines worth quoting. Pick one strong idea from the hour rather than summarising the whole thing, and keep it to eight hundred to fifteen hundred words.
How do I transcribe a podcast episode without uploading it?
Use a tool that transcribes on your own device instead of sending the file to a server. Yaps Studio imports your audio or video file and transcribes it locally on Windows, macOS, or Linux, so an unreleased episode never goes to a third party and there is no upload wait before the work starts. Most per-minute transcription services have to receive the file before they can process it.
What is the best podcast notes app for turning episodes into articles?
Yaps is the strongest choice when the recording should stay private and you publish often, because it transcribes on your own machine, labels speakers on Windows and macOS, and lets you dictate the write-up in any app. Transcripts and drafts live in local notes with local search, so your whole archive stays findable. If a one-click push to your podcast host matters more than keeping the audio local, a cloud service with that integration is the better fit for that specific need.
Can Yaps tell me who said what in a podcast interview?
Yes, on desktop Windows and macOS, and it is optional rather than automatic. Install the speaker identification pack from the Features screen in the app, a one-time download of around 32 MB, then switch on "Identify speakers" before you start the transcription. It is off by default and it is not available on Android or Linux, and you should still skim the result and correct the occasional label.
How do I write show notes from a transcript?
Read the transcript once, then dictate five or six sentences on what the episode covered and a list of the topics in the order they came up. The topic list is the part listeners actually use, because it tells them whether the episode is for them. Add any names, links, or references mentioned, which the transcript makes easy to find, and keep the whole thing short enough to read in a feed.
Does Yaps record my podcast?
No. Yaps does not record your podcast, does not join calls, and is not a remote recording studio. You record with your own setup and bring the finished audio or video file to Studio afterwards, which then transcribes it on your machine. The workflow starts with a file you already own.
How long does it take to turn an episode into an article?
A little over an hour of your attention for an hour-long episode, once you have done it twice. Roughly fifteen minutes goes on reading and correcting the transcript, five on dictating show notes, twenty on dictating the article draft, and twenty-five on editing, pulling quotes, and exporting, which comes to about sixty-five minutes in total. The transcription itself is hands-off, so you can leave it running.
Can I get SRT captions for the video version of my podcast?
Yes. Studio exports SRT subtitle files from the same transcription job, so a video cut of the episode can go out with captions. The export formats are WAV for audio and SRT for subtitles, not MP3 and not VTT. Uploading the SRT alongside your video is a manual step, because Yaps does not connect to any video platform.
Does Yaps publish the article to my podcast host or blog?
No. Yaps has no integration with any podcast host, content management system, or social scheduler. It produces clean, paste-ready text and exported files, including Markdown, plain text, WAV, and SRT, and you paste or upload them yourself. That last step takes about a minute per asset.
How much does it cost to transcribe a podcast with Yaps?
Studio is a desktop feature, and desktop Yaps starts with a 7-day free trial and then costs $15 a month for Pro, or $12 a month billed annually. There is no permanently free desktop tier, and starting that trial needs an account, which governs access and billing rather than where the work happens, because the transcription still runs on your own machine. Because it is a flat subscription rather than a per-minute meter, a ninety-minute episode costs the same as a fifteen-minute one, which is what makes transcribing an entire back catalogue reasonable. The Android app is free and needs no account, though file transcription is desktop only.
Can I do this workflow on my phone?
Not the transcription part. Studio file transcription and speaker identification are desktop features on Windows, macOS, and Linux. The Android app is free and handles voice typing in any app, voice notes with local search, and read-aloud, so you can dictate show notes or an article draft on a phone, and vault sync can carry notes between mobile and desktop on a paid plan. An iPhone and iPad version of Yaps is coming soon.
Does the transcript work if the episode is in another language?
Broadly, yes. Yaps speech recognition understands about 25 spoken languages and detects the language automatically from the audio, so a non-English episode is supported in general. For a specific show in a specific language, the sensible move is to test a short clip during the trial before committing a long archive to it.
Is transcribing offline cheaper than paying per minute?
For anyone publishing regularly, yes, and the bigger effect is on what you choose to transcribe. A per-minute meter pushes you to triage, so the small episodes and the back catalogue never get done. A flat subscription running on hardware you already own removes the per-episode decision entirely, which is how you end up with a searchable transcript of every show you have ever made rather than the three you judged worth the spend.