Seedance 2.0 Fast Free Online AI Video Generator

Nine Reference Images, as on Standard 2.0

Fast does not trim the inputs to save time. Its composer takes the full Seedance 2.0 reference set, up to 9 images and 3 clips, and every upload gets a tag such as @Image 2 that you write into the prompt to say what it is for: the person, the outfit, the vehicle, the light.

The demo below is ByteDance’s own four-image example, rendered by standard Seedance 2.0; no Fast sample of it has been published. Fast accepts the same four pictures and the same prompt, so this is the run you would start, with a shorter wait and a lower charge.

Input

1

2

3

4

Seedance 2.0

Portraits and a Motion Clip in One Run

Some actions are easier to show than to describe. Attach a clip of the movement, two to fifteen seconds long, next to the portraits of your characters, and tell the model which file does which job. Up to three clips can sit in one run.

A reference clip changes the bill: Yovid counts the clip’s seconds plus the output’s seconds, at a lower per-second rate than a plain run. Reference images never add to the charge. This demo, again rendered by standard Seedance 2.0, borrows a slurping action from a clay-figure clip for a two-person scene.

Input

▶ 1

1

2

Seedance 2.0

Put a New Product into Footage You Have

A reference clip can be the whole shot, not just its motion. Here one image of new packaging replaced every bag in a 15-second pet-food spot, through camera moves and partial cover, while the dog, the room and the cuts stayed as they were. Standard Seedance 2.0 made this example; Fast takes the same clip-plus-image setup.

Fast renders a new clip from the reference rather than editing the file in place. For changes to an uploaded video that keep everything else frame for frame, or to continue a clip past its end, Seedance 2.5 has dedicated edit and extend modes.

Input

▶ 1

1

Seedance 2.0

Shorter Waits at 480p and 720p

The progress bar in Yovid’s generator runs on median submit-to-finish times for jobs that did not wait in a queue. By that table Fast saves about a third of the wait at 480p and nearly half at 720p. At 1080p the two models take about the same time, so the saving there is in credits rather than minutes.

On Pro and Premium the ×N switch beside Generate starts up to 4 or 8 takes of one prompt at once, which pairs well with a model built for trying variations.

Fewer Credits per Second than Standard 2.0

Fast costs less per second than standard Seedance 2.0 at 480p, 720p and 1080p, and sound is included either way. The figures below are credit-pack rates; the composer shows the exact charge for your settings before you press Generate.

From a Picture, Fast Stops at 720p

Attach an opening picture, and a closing one if the shot must land somewhere set, and the resolution menu narrows to 480p and 720p. Text-only and reference runs keep 1080p.

When a still has to become a 1080p clip, standard Seedance 2.0 keeps every resolution from a first frame, up to 4K, and Seedance 1.5 Pro does 1080p from a frame at a much lower rate.

All the Settings, Side by Side

The composer opens on 16:9, 720p and 5 seconds. Adaptive lets the model choose the shape from the prompt or the input picture; pick a fixed shape when the clip has to fit a placement.