Pixverse Transition: the still is the contract (5s, 8s, 10s, 1080p, 40cr)

Pixverse Transition is image-to-video. If the label is wrong, every second is waste.

Pixverse Transition creates smooth transitions between video clips. Category in the catalog is edit-video. requires_image is true. audio is null. Durations are 5s, 8s, and 10s. max_output_resolution is 1080p. Forty credits is the 1080p call. The stills are not optional flavour. They are the two ends of the join.

Two frames. One join. Three lengths.

This row is not a scene generator. It is not "a cinematic shot of a bottle." It takes a first frame and a last frame and spends 5, 8, or 10 seconds travelling between them. If either still is the wrong SKU, the wrong face, or the wrong room, every second of that travel is waste. You cannot prompt the label onto a bottle the first frame does not show.

requires_image is the gate. Text-only will not run. Approve both stills on a text to image pass — or shoot them — before you open this row. Lock the still is the same rule; here it applies twice.

The AI image-to-video tool is the Versely door for animating a plate. Pixverse Transition is the join-shaped version of that job: two plates, one interpolate. Pixverse's other rows share the bill. Best Pixverse model is the family shortlist. This page is only the transition.

1080p, no soundtrack from this row

Quality settings include 360p, 540p, 720p, and 1080p. The 1080p option is the 40-credit figure in the catalog matrix. Do not run 1080p to hide a soft still. Upscale or regenerate the stills; do not ask a transition to invent grain that was never in the frames.

audio is null. This join does not come with native dialogue, room tone, or a score. Plan music, TTS, or silence. Add music after picture lock if the cut needs a bed. Do not write "whoosh and a voiceover" into the prompt and expect a stem.

Aspect on the row: 16:9, 9:16, 1:1, 4:3, 3:4. Both stills should already be that ratio. A 16:9 first frame into a 9:16 last frame is a different crop, not a transition. Match the files, then pick the duration.

If the label is "scene" you are already wrong

A scene brief is text-to-video. A single-plate push-in is ordinary image-to-video. A morph between two approved frames is this row. Routing a scene brief here produces a confident interpolation between two pictures that were never a shot list.

Write the prompt as the travel: what should stay constant (the product, the face) and what is allowed to change (angle, time of day, wardrobe) between frame A and frame B. Do not also recast the subject. The stills already cast it.

Five seconds is a snap. Eight is a hold. Ten is the long interpolate. The catalog does not offer 4s or 15s on this slug. If you needed a 15-second world, you are on a different model. Stretching a 10s join in the timeline is a retiming artefact, not more Pixverse.

Forty credits at 1080p is one join. It is not a pack of scene takes. If the first still is wrong, stop. Edit the still. Do not reroll the transition hoping the bottle becomes the right bottle.

FAQ

Does Pixverse Transition generate audio?

No. audio is null on this row. You get a silent 5s, 8s, or 10s interpolate at up to 1080p. Score or VO afterwards. Do not treat a mute file as a bug.

Can I run it from a prompt with no images?

No. requires_image is true. The job is a join between frames you supply. A prompt without stills is a different category.

Why 40 credits instead of a cheaper video row?

That is the 1080p catalog figure for this slug (360p and 540p list at 15, 720p at 20 in the same matrix). You are paying for a transition between two plates, not for a text-to-video sample. If you did not need the join, do not buy it.

Is this the same as first-last-frame on other families?

Same job family: two stills, a path between them. This slug is Pixverse, edit-video, 5s / 8s / 10s, 1080p, 40 credits at the top tier, no audio. Copying another vendor's duration list onto this row will 400 in your head before it 400s in the API.