01 — Check the rights before the file
Work out what you are actually allowed to publish
If the channel is yours and every voice on it belongs to you or to a guest who agreed to be published, you can move ahead. That covers most people arriving at this page.
Two things trip people up. The first is background music. A licence that covers a YouTube upload very often does not cover a podcast feed, and music libraries treat the two as separate uses. Check the licence terms before you publish rather than after a takedown.
The second is guests. Someone who agreed to appear on your channel has not automatically agreed to appear on a podcast. A short message asking is quicker than the conversation you will otherwise have later.
Someone else's video is not yours to republish, whatever tool makes it technically easy.
If the music licence is the only blocker, it is usually easier to strip the music and re-score the audio version than to renegotiate the licence.
02 — Get the audio out of the video
Start from the best source you have
Use your original recording if you still have it. The file that went to YouTube has been compressed for streaming, and starting from the master gives you noticeably more room when you clean the audio up.
If the original is gone, download your own video from YouTube Studio and pull the audio track from it. Upload the file, take the audio out, and you have something you can edit.
Expect the extracted audio to need more work than a recording you made for audio in the first place. Video is forgiving of a room that audio is not, because a viewer watching your face will tolerate an echo a listener in headphones will not.
03 — Fix everything that pointed at the screen
The listener cannot see what you are showing
This is the step that decides whether the episode works. Every "as you can see here", every "this chart", every silence while you clicked through something, becomes a hole in the audio.
Read the transcript and mark each one. You then have three options for each: cut the passage entirely, record a short replacement line that describes what was on screen, or keep it and accept the listener will miss something. Cutting is usually right for a screen walkthrough. A replacement line is usually right for a single chart or number.
Record the replacement lines in one sitting with the same microphone setup, then drop them into place. A few seconds of new audio can rescue a section you would otherwise lose.
- Every "as you can see" is cut or replaced
- No unexplained gaps where a visual was on screen
- Charts and numbers are spoken aloud
- Anything that only made sense on screen is gone
04 — Tighten it for a different kind of attention
Video pacing is slower than audio pacing
A viewer has a picture to hold their attention through a slow passage. A listener on a commute has nothing. The pauses and repetitions that felt natural on camera tend to feel like dead air in headphones.
Cut the on-camera housekeeping first: the subscribe reminder, the wave at the start, the comment prompt. None of it means anything on a podcast feed. Then tighten long pauses and remove the places where you circled back to the same point.
Editing against the transcript rather than the waveform makes this much faster, because repetition is obvious in text and almost invisible in a waveform.
05 — Give it an opening that works without a picture
The first fifteen seconds carry more weight
A YouTube viewer already saw a thumbnail and a title before pressing play. A podcast listener may have queued this from a feed with none of that context.
Record a short new opening that says what the episode covers and who is speaking. Fifteen to twenty seconds is enough. If there is a guest, name them and say why they are worth listening to.
Level the whole episode at the end so the new opening does not sit noticeably louder or quieter than the extracted audio behind it.
06 — Publish it to a feed and point the two at each other
Two channels, one piece of work
Publish to a podcast feed and submit it to the directories once. After that every new episode flows through automatically and you are not repeating the setup.
Put the transcript in the show notes. You already produced it while editing, it costs nothing to publish, and it gives the episode something to be found by.
Then link both ways. Mention the feed in the video description, and point listeners at the video when a passage was genuinely visual. The two versions reach different people and there is little overlap between them.
Pull the audio and publish it as an episode
Extract, edit against the transcript, and publish to a feed in the browser.
One recording, more than one channel
Explore the complete video-to-audio workflow.