Universal captions and subtitles

Every cue, directly addressable.

A lossless, structured interchange layer
for subtitle and caption content.

Cue JSON

3 cues

00:18:04.120
00:18:06.980cue_0042

The archive keeps the original frame.speaker: narrator · region: lower-third

OCR 0.97

00:18:07.440
00:18:10.220cue_0043

The words become a predictable data surface.speaker: narrator · source: bitmap subtitle

OCR 0.94

00:18:11.000
00:18:13.640cue_0044

Downstream tools read one Cue JSON shape.speaker: narrator · provenance attached

OCR 0.99

Readable text and timing travel with the original graphics, metadata, layout, and provenance.