Descript captions: in/out animations vs Sume design overrides

Descript's September 17 release added caption presets and text animations with separate in and out motion. Sume names no animation catalog, but its video-captions API lets you set the timing of that motion per request: motion.enter_seconds and motion.exit_seconds, plus emphasis_in_seconds and emphasis_out_seconds for the spoken word.

Descript facts are from its changelog; Sume facts are from Video captions, both read 2026-09-30.

What did Descript change in captions?

The changelog entry for September 17, 2026 lists 18 new caption presets from Descript's design team, and 10 new text animations for 15 total, each with separate in and out motion. It also lists search and filtering in the preset list and a rebuilt properties panel. Presets apply when you click them, and existing styles stay in the list.

What does Sume let me control on a caption's motion?

In Sume a style is a set of design tokens, and the optional design object overrides them for one request. Each field merges over the style's own value, so one key changes one thing. The motion group has four timing fields; phrasing controls how words are grouped.

Is that the same as choosing an animation?

No. The docs describe timing tokens on a named style; they do not list a catalog of animation types. The look and the motion come from the style you name, and design edits it. Colours are hex, rgb()/rgba() or transparent, and numbers outside their documented range return 400, so a bad value fails at request time instead of rendering wrong and billing.

One limit to know: design is not supported on punch or tiktok-green, which render on a path that reads none of these tokens.

How do I shorten the entrance of a caption?

Send a design object with only the key you want to change, to POST /v1/video-captions with a public HTTPS video_url. The docs' own example overrides a single colour; the same shape applies to motion. See Customize captions API for the colour, placement and size tokens.