Your Podcast Went Video – Key Things to Understand
AUGUST 2026 – Apple switched on video podcasts earlier this year. Spotify opened its Distribution API shortly after. YouTube has quietly been the biggest podcast platform for years and finally stopped being coy about it.
For fifteen years the deliverable was an MP3 and a paragraph. Now it's eight assets and three platforms that each want something slightly different.
Here's the part every roundup skipped: the platforms were ready long before the hosts were. Video doesn't move through RSS anymore. It moves through your hosting company. So for most of this year, whether you could publish video had almost nothing to do with Apple and almost everything to do with who you pay for hosting.
That gap is why we moved Merrick's shows to Buzzsprout. More on that below, because it's the most useful thing in this post.
Apple: real video, real gatekeeping
Apple's rollout is the genuine change, and it does not work the way podcasting has always worked.
Video is delivered as an HLS stream, pushed by your host through a private ingest API. It does not travel in your feed. Four hosts had access at launch in February. Ten by March 16.
Which means the honest headline: if you self-host, you cannot publish video to Apple. Not with a workaround, not with a clever feed tag. The door is closed and Apple built it that way on purpose.
It plays on iOS, iPadOS, visionOS 26.4 and the web player. Not Apple TV, which surprises everyone. Paid subscriptions stayed audio only.
One thing that catches producers off guard: once an HLS stream is attached to an episode, its audio track becomes the only audio Apple listeners hear. Your mastered MP3 stops reaching them. Whatever you bounced into the video is now the show. Master it accordingly.
Spotify: the money moved
Spotify went the other direction and published actual specs, which after Apple feels like a gift.
MP4, H.264 or H.265. Up to 60GB and twelve hours. 1080p or better, 25 Mbit/s CBR. Sixteen by nine or you get letterboxed, and letterboxing reads as amateur faster than bad audio does.
One warning with teeth: do not revert a Spotify feed to audio-only after you've uploaded video. Every video goes with it. Not archived, not recoverable, gone.
YouTube: pick a lane
Point your RSS feed at YouTube and it generates a static-image video from your show art. Free reach, zero work, and the audio locks once it publishes. Distribution is YouTube-only.
Upload natively and you get custom thumbnails, chapters, retention analytics and Partner Program eligibility. Everything that makes YouTube a growth channel instead of a mirror.
If you're producing video anyway, the difference between those paths is about five minutes of effort and roughly an order of magnitude of result.
Why we moved to Buzzsprout
Now the practical part.
Through most of 2026, "supports Apple video" and "you can publish Apple video" were different sentences. Hosts announced partnerships and then ran waitlists. Some required you to generate an API key in Apple Podcasts Connect and request approval from Apple directly, with a stated turnaround of up to two weeks. Others reserved it for enterprise accounts. One major host rolled out gradually and had a few dozen shows live months after announcing.
Buzzsprout turned Apple video on for everyone on May 4, and was reported as the first host to make it generally available at all. No queue, no invite, no enterprise tier.
Then on August 5 they finished the job: video publishing to Spotify and YouTube, available to every Buzzsprout podcaster. Upload one file, publish HD video to Apple, Spotify and YouTube, keep distributing audio everywhere else. Same workflow you were already using.
That last part is the whole argument. Everywhere else, Apple and Spotify are two separate pipes that don't talk to each other, and you publish twice. Buzzsprout collapsed that into one upload. For a weekly show, that's not a convenience. That's an entire recurring task off the board.
Is it perfect? No. Video plans start around $30 a month with a monthly cap on video hours, so a daily long-form show should check the math before committing. And you're trusting one company's roadmap across three platforms.
We took that trade, with some hesitation, and at this point we'd recommend it to any show that isn't already locked into an enterprise contract. Right now Buzzsprout is the best solution on the market for publishing a video podcast, and it isn't close. They shipped first, they shipped to everybody, and they're the only ones who made the whole thing one step.
If you're on another host, ask them one specific question. Not "do you support Apple video." Ask: can my show publish video to Apple, Spotify and YouTube this week, from one upload. The answers get very short.
The delivery list
Eight things. Assume all of them, every episode.
Video master. MP4, 16:9, 1080p or better, H.264 or H.265. One file feeds everything.
Audio master. Still required, still mastered to -16 LKFS. RSS runs on it and Apple's paid tiers are still audio only.
VTT captions with speaker names. The one people skip. Apple's own documentation says providing a VTT lets you identify every speaker with each line. SRT can't do that.
SRT captions. The fallback. Generate it from the VTT, never the other way around.
Thumbnail. Sixteen by nine, and not your show art. Show art is a square logo. A thumbnail is a separate 16x9 image.
Chapters. Apple's rules are the strictest published: at least three, first at 00:00:00, no more than six per hour, each running at least two minutes, nothing under a ten-minute episode. Titles under 45 characters and five words. Meet Apple and you've met YouTube.
Copy block. Title, short description, long description, tags to 500 characters.
Vertical clips. Fifteen to ninety seconds. Six is healthy for social. Spotify takes exactly one per episode, so choose well.
What nobody has published
This is where most advice online is confidently wrong.
Apple has published no video encoding spec. No resolution, no aspect ratio, no codec, no bitrate, no maximum duration. Every number you've seen quoted was reverse-engineered by someone reading a delivered stream. Useful. Not authoritative.
Apple says a VTT identifies your speakers but never documents the syntax it expects. The standard voice-span tag is the reasonable bet, not a confirmed one.
Apple's docs say chapters render on video episodes. Testing earlier this year found them on neither the scrub bar nor the player, and that held even when the listener switched to audio-only. On audio-first episodes they render fine. Build them to spec, expect them to vanish on video.
Spotify hasn't published a podcast loudness target, and its minimum audio bitrate is quoted differently depending on where you look.
None of this is a reason to wait. It's a reason to test on a real episode before you build a workflow around a number you found in a forum post.
Six things worth knowing
Cut for 16:9, frame for vertical. Keep faces in the center third. Every clip becomes a crop instead of a re-edit, and six clips takes an afternoon instead of a week.
Write the VTT first. It's the superset. Speaker names live there. Derive the SRT from it and you never retype a name.
Own the transcript where it compounds. Leave Spotify's alone. Machine captions mangle names, brands and jargon, so write your own for Apple, for YouTube and for your website, where Google actually crawls it. Spotify is the exception. You can disable transcripts in Spotify for Creators, and you shouldn't. An imperfect transcript still serves deaf and hard-of-hearing listeners, and auto errors cluster in proper nouns, which bruise your ego more than they hurt anyone's comprehension. Turning it off costs you an audience and buys you nothing.
Two Spotify transcript traps. You can upload your own and it overrides theirs, but the controls are not available to every account. And if you upload a transcript and then switch auto-generation on, your upload gets replaced. Worse, Spotify has not committed to ingesting the podcast:transcript tag from third-party host feeds, so a perfect VTT sitting in your RSS may never reach Spotify at all. If you want your transcript there, upload it in Spotify for Creators, if your account allows it (nothing written why some accounts can and others can not) and leave auto-generation off.
Pull clips from the transcript, not the timeline. Reading an hour is fast. Scrubbing it isn't. Find the line, take the timestamp, then open the edit.
Fewer chapters, better placed. Six an hour is a ceiling, not a target, and since iOS 26.2 Apple auto-generates a list for most English episodes anyway. Yours only earns its place by being better than the machine's.
Keep mastering the audio. Video is a new surface, not the only one. A large share of your audience is still driving.
The actual takeaway
The barrier to podcasting just went up, and that's the opportunity.
An MP3 and a paragraph was a low bar. Anyone could clear it, and roughly everyone did. Eight assets across three platforms is a real production standard, and standards separate shows that are built from shows that are posted.
The good news is the hard part just got solved. Six months of this was a hosting problem, and as of August it's one upload. Get your assets right, get on a host that can actually ship them, and you're competing on the work.
Merrick Creative and Merrick Studios build production systems for shows that ship on schedule. We're on Buzzsprout because we tested the alternatives and it wasn't a close call. No affiliate deal, no sponsorship, just the thing we actually use. Specs current as of August 2026.
