What can the Compose API do to a video?
117 features across the timeline, media, text, captions, graphics, layout, motion, transitions, picture and sound effects, and output. Every one is a field in the same JSON timeline, and 80 of them have a rendered preview here.
How do I add an effect to a clip in a compose job?
Give the clip an effects list and put one entry in it per effect, each with a type and only the settings you want to change. The list runs in the order you write it, and the same list works on a track or the whole timeline.
Do effects, transitions and titles cost extra?
No. Compose bills one rate per minute of finished video, $0.04/min, however many features the timeline uses. Animated graphics such as kinetic titles and 3D turns render up to 1080p at 30 fps on self-serve plans.
Can I animate a title word by word?
Yes. Give the text an animate entrance with by set to letter, word or line, and add out to animate it leaving. Stagger and order set the rhythm, so a kinetic title is a few fields on one text clip.
How do I check a timeline before I render it?
Send it to POST /compose/validate, or call compose.validate in the SDK. It parses the timeline exactly as a render would, costs nothing, and returns any problems, the warnings and what the render will cost.
Can an AI agent use these features?
Yes. Through the Rendobar MCP server, Claude, ChatGPT and other agents read the same list of effects, check a timeline with validate_compose, and render it. Every preview here is a request an agent can copy as written.