9-second Kling motion control clip: $1.42, or $1.64 finished

A 9-second Kling 3.0 Motion Control clip costs $1.4175 on Sume, which is 9 x $0.1575. Muting the original sound and adding a Music 1.0 track ($0.125) and a one-minute timeline render ($0.10) brings the finished clip to $1.6425 (as of 2026-10-09).

The line items

Motion control takes a still image (or ready avatar) and a motion reference video, and the reference video drives all movement in the output. The price is per second of output, taken from the declared duration_seconds (1 to 30) rounded up. A 9-second reference is exactly 9 billable seconds.

Why the music step exists

The motion route keeps the reference video's sound by default. Setting keep_original_sound to false returns a silent clip, which is the option to use when the reference contains someone else's voice or a track you do not have rights to. A silent clip then needs a sound bed, and the timeline renders the clip and the bed into one MP4. Timeline pricing is $0.10 per output minute with a one-minute minimum and a 30-minute maximum, so a 9-second clip pays for a full minute.

Ten clips a week

At ten such clips, the motion line is 10 x 1.4175 = $14.175, the music is $1.25 and the renders are $1.00, for $16.425 in total. The music line is the easiest to cut: reuse one track by joining several clips in a single timeline render, which bills per output minute, so ten 9-second clips joined into 90 seconds are one render of 2 billable minutes, $0.20.

Check the timeline docs for the audio spine and video slot rules before you plan the join. Each motion clip has to be a Sume-hosted video for the timeline to take it, and the result artifacts are on media.sume.com.

Using it through MCP

The MCP tool for this route is kling-motion-control_create. The tools page lists it among the paid generation tools, which means it needs the write grant at OAuth consent; there is no separate paid scope. See MCP tools and gates. With an API key you call the route directly instead.

For a first test, keep the whole thing under $1.50 by skipping music and render and watching the silent clip; you will know within one job whether the still and the reference suit each other.

For longer pieces, remember the 30-second ceiling. A 30-second reference is 30 x 0.1575 = $4.725, and two 15-second references cost the same $4.725 in total, so splitting a long reference only makes sense when you want separate clips.

• Cheapest 9-second motion test: keep the reference at 9 seconds and skip the music step (1.4175).
• A 10-second reference costs 10 x 0.1575 = $1.575.
• The 30-second ceiling is 30 x 0.1575 = $4.725.