Instagram auto captions vs burned-in captions: when to burn them in

Use Instagram's auto captions when you post the Reel inside Instagram and want to edit the words there; burn captions into the video when the same MP4 will also go to other places, or when the clip has no speech to transcribe. Instagram's guide says captions are worth including even if you already added text overlays, and Sume video captions is the burn-in route.

Instagram's side comes from its Reels guide and the creator editing tips, both read on 2026-10-03. Sume's side comes from the video captions docs page.

What do Instagram's own captions do?

According to the guide, Instagram automatically generates captions from the audio. You can review and edit the transcribed text, change the caption style (font, color, background) and reposition the captions on screen. The creator tips add that on-screen text keeps viewers engaged, especially when they watch without sound.

The guide does not say whether Instagram will add its caption layer on top of captions that are already burned into the picture, so test one post before you run a batch. It also does not give a caption length or a safe-zone number.

When is a burned-in caption the better choice?

Burned-in captions are part of the pixels. That has costs and benefits, and the table keeps them side by side.

How do I burn captions with Sume?

Send video_url and, if you want, a style. Omit the style and the wording decides: slam for Latin text, black-outline for Korean. You can name slam, punch or tiktok-green for Latin copy. language is only a speech-to-text hint, so it never picks the look.

If you already have the script, pass script_text. Sume keeps the speech-to-text word timings and aligns your wording to them; if alignment fails you get script_alignment_mismatch or script_alignment_failed, and the suggested fix is to simplify the script or omit it.

curl -X POST https://api.sume.com/v1/video-captions \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: reel-caption-001" \
  -d '{
    "video_url": "https://media.sume.com/artifacts/example/clean.mp4",
    "style": "slam",
    "language": "en"
  }'

What about a silent clip or a restyle?

A silent clip fails as caption_no_speech with next_action: use_overlay_captions. For that case pass cues (or segments), each with text, start and end in seconds; the docs say that skips speech-to-text and burns exactly that copy at those times. The fields script_text, words, cues and segments are mutually exclusive.

To try a different look without transcribing twice, send source_caption_id instead of video_url. Sume reuses that caption's source video and word timings, and billing is unchanged because a restyle is still a render. Poll with GET /v1/jobs/:id/status and read the result as described in Jobs and results.

What should I check before I post?

Play the burned file on a phone at normal brightness and read the lines without sound. Confirm that no caption sits under Instagram's own interface elements; Sume's design.placement.anchor_ratio moves the line, but Instagram does not publish a safe-zone number in the guide, so judge it on your own screen.

Then post one test Reel with burned captions and check whether Instagram's auto captions double up. If they do, turn the in-app captions off for that post. Sume cannot control that setting for you.