YouTube podcast-to-Shorts conversion: four countries, or Sume

YouTube's audio-first video conversion, which turns podcasts into Shorts in YouTube Studio, is available only in Australia, Canada, the UK and the US, according to YouTube's help page for AI features in Shorts. Anywhere else, you cut the clips yourself. With a video podcast recording, Sume can find, cut and caption them by API.

The YouTube facts here are from that help page and the disclosure page, read 2026-10-02.

What does YouTube's audio-first conversion do?

The page lists "Audio-First Video Conversion" as turning podcasts into Shorts in YouTube Studio, and notes the availability above. It generally requires the YouTube device language to be English. The page does not describe how clips are chosen, how long they run, or whether you can edit them, so I make no claims about that.

On disclosure, the page says YouTube's own generative AI tools disclose use automatically. It does not say whether this conversion counts as one of those tools, so check the disclosure screen before you publish.

How do I cut podcast Shorts myself?

Sume's routes work on video files, not bare audio, so this needs a video recording of the podcast: a camera recording, a studio recording, or a static-image video exported from your editor. The recording must be a media.sume.com clip, so import it first with POST /v1/media-imports.

The workflow has three jobs.

How do I get a transcript to pick the moment?

Video inspect with transcribe: true returns the transcript with words and, if you set segmentation.mode to sentence, gapless sentence segments. You read them yourself and choose the start and end; Sume does not pick a "best moment" in this route. Check probe.has_audio first, since a clip without audio fails with inspect_source_has_no_audio.

Ask for sentence segments on a 20-minute episode:

curl -X POST https://api.sume.com/v1/video-inspect \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: pod-ep12-inspect" \
  -d '{
    "video_url": "https://media.sume.com/artifacts/artf_demo/ep12.mp4",
    "frames": false,
    "transcribe": true,
    "duration_seconds": 600,
    "segmentation": {"mode": "sentence"}
  }'

How do I cut and caption the clip?

Pass the chosen start and end to video trim, then pass the result's video_url to video captions. Captions need audible speech; a silent clip fails as caption_no_speech. Note that the transcript duration_seconds hint maxes at 600, so for longer episodes inspect in sections.

Because a podcast Short is mostly a face and words, burned-in captions matter. YouTube's own auto-captions are limited in languages; see auto-captions not showing for the fallback.

What does this cost, and what is missing?

For one 45-second clip from an episode: $0.02 for the trim, plus the transcript at $0.01 per audio minute, plus the caption job at the live catalog price. Sume does not rank moments, write titles or upload to YouTube. If you want the longer webinar-style pipeline, see Webinar recording to shorts.

How long should a podcast Short be?

The page for the audio-first feature gives no length. The general Shorts limit is three minutes. For a talking-head clip, the transcript gives you natural sentence boundaries to cut on, so you can end on a full thought rather than a fixed second count. Cut a little early rather than late; trailing silence costs attention.