Add a lower third name tag to a video with an API

Sume has no lower-third template, but you can burn a name tag with the video captions endpoint: send one authored cue whose text has a newline (name on line one, title on line two), with start and end times, and place it with design.placement.anchor_ratio. Authored cues skip speech-to-text, so this works on a clip with no speech.

Steps are from Sume's Video captions page and OpenAPI schema, read 2026-10-01. Opus Clip's changelog (Sept 28) is the trigger: it added lower thirds with Name Tag and Location presets.

What does a name tag cue look like?

The create call is POST /v1/video-captions with a required video_url. The schema describes cue text as one that "may include a newline for a 2-line card". Pass cues and Sume burns the overlay without ASR.

{
  "video_url": "https://example.com/interview.mp4",
  "cues": [
    { "text": "Dana Lee\nHead of Product", "start": 1.0, "end": 5.0 }
  ],
  "style": "black-outline",
  "design": { "placement": { "anchor_ratio": 0.82 } }
}

How do I control where the card sits?

design.placement.anchor_ratio is the line centre as a fraction of frame height; landscape_anchor_ratio does the same for landscape. A value near the bottom gives a lower-third position. Check the result on your own footage, since the docs give the parameter's meaning but no recommended number.

What can this not do?

It is a styled caption, not an editable lower-third object: no animated bar graphics, no Location preset, and no 12 presets as in Opus Clip. design overrides are not supported on the punch and tiktok-green styles, and Korean text on a Latin style such as slam returns a 400, so use black-outline or another Hangul style for Korean names.

Can I add more than one name tag?

Send several cues, each with its own start and end, one per speaker. For other burned-in copy on silent clips, see authored overlay cues.