Timeline slots must reach within 0.5 s of the voiceover: 24 s example

In Sume's Timeline 1.0, the video slots must start at 0 and run to within 0.5 seconds of the end of the audio. For a 24-second voiceover, four 6-second slots starting at 0, 6, 12 and 18 seconds cover it exactly, and the render costs $0.10 because it is one started output minute. Use the free plan endpoint to confirm the layout before you render.

The rules below are from the Timeline 1.0 docs (read 2026-10-06). The audio sets the length of the output, and the video slots are laid on top of it.

The layout rules

Each slot declares its own start on the audio spine, a duration of at least 0.2 seconds and, optionally, a source_in point into its file. Declared starts are authoritative, and the compiler compensates for any crossfade, so you never overlap slots by hand to make room for a transition.

Plan before you pay

POST /v1/timeline-1.0/plan runs the same schema checks and the compiler without creating a job or reserving credits. It returns the duration, the segment count, the billable minutes and the estimated cost, plus a summary of the filtergraph. It cannot predict warnings about short sources being padded or looped, so a clip that is 4 seconds long in a 6-second slot will only show that warning on the real render.

Every URL must already be a media.sume.com artifact or asset in your workspace, so import files first. The render needs an Idempotency-Key, and the default output is 1080 by 1920. Set the output size if you need 16:9.

When it goes wrong

The common mistake is a short last slot: a 23-second sum of slots against 24 seconds of audio is inside the 0.5 second allowance only if it ends at 23.5 or later. Extend the last slot, or trim the audio spine with source_in.

A second mistake is leaning on transitions to hide gaps. Transitions are limited to one second, and Sume refuses more than eight chained fades. For a four-slot ad, use a hard cut or a short fade between slots and put the effort into the clips.