YouTube Shorts Clear screen mode: check your burned-in text

Does Clear screen mode hide captions burned into my Short?

YouTube describes Clear screen as a viewing option that hides all icons and text for a distraction-free view of a Short. That reads as the player's own icons and text. Captions burned into your MP4 are pixels of the video itself, so I expect them to stay on screen, but YouTube's announcement does not say so in as many words. Treat it as a test to run on a real phone, not a fact.

The practical rule does not depend on the answer: anything a viewer must read has to live inside the video file, because the like, comment and share controls can disappear in this mode and your description is not on screen at all.

What YouTube announced for the Shorts player

Clear screen arrived together with other player changes. The page says the updates are rolling out over time starting June 25, 2026, with no completion date given.

Pull stills and look at what the viewer would see

Sume's video frames route extracts up to 24 stills from one clip at times you name, at the source size, as durable image artifacts. The source must already be on media.sume.com, so import it first, and the source can be up to 300 seconds. Ask for PNG if you want to read small text without compression noise.

Pick times just after each caption starts and at the last second, then open the stills side by side. You are checking that nothing important sits in the corners where player icons normally appear, and that the call to action is not a hand-written line in the description.

curl -X POST https://api.sume.com/v1/video-frames \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: clear-screen-check-001" \
  -d '{
    "video_url": "https://media.sume.com/artifacts/artf_demo/short.mp4",
    "at": [0.5, 3, 6, 9, 12],
    "format": "png"
  }'

Burn the words you cannot afford to lose

If a key line currently lives only in a text sticker you add in YouTube's app, move it into the file. Sume's video captions route burns text onto a finished clip. Passing cues with text, start and end burns exactly that copy at those times and skips speech-to-text, which also makes it the path for a clip with no speech. A standalone caption job is priced at $0.20 for videos up to 60 seconds under the current fixed estimate, and the docs say to confirm live pricing in the catalog.

The design object can override placement, phrasing and colours for one request, so you can keep lines short and away from the edges where controls normally sit. Then run the frames check again on the captioned output.

curl -X POST https://api.sume.com/v1/video-captions \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: clear-screen-captions-001" \
  -d '{
    "video_url": "https://media.sume.com/artifacts/artf_demo/short.mp4",
    "style": "slam",
    "cues": [
      {"text": "Free shipping ends Friday", "start": 0.5, "end": 3.0},
      {"text": "Code SPRING at checkout", "start": 3.0, "end": 6.0}
    ]
  }'

A short checklist before you upload

Run these in order on the exact file you plan to publish.

• Confirm the offer, price or call to action is readable in a still taken from the final MP4, not only in the description.
• Check the stills at the start, middle and end for text that collides with where player controls normally sit.
• Watch the upload on a phone once with Clear screen turned on and note what disappears.
• Keep the stills with the job id so a client can see what was checked.

If a still shows a collision

Fix it at the source of the problem rather than in the upload screen. If the words sit too close to an edge, re-burn the captions with a different design and compare the new stills. When you pass source_caption_id instead of video_url, the docs say Sume reuses the stored source video and the word timings it already has, so no second speech-to-text run happens, and billing is unchanged because a restyle is still a render.

Keep each revision's job id next to its stills. When a client asks why the final version differs from the first, you have the comparison.