PrunaAI P-Video: 2 credits per second, up to 20s

PrunaAI P-Video is a published text-to-video and image-to-video row at 2 credits per second of output, 1080p, 1 to 20 seconds. It is a low per-second draft row. It is not Seedance 2.5 (from 34 credits, native audio), and it is not a 2-credit background remover.

PrunaAI P-Video is a generator. You did not already have the clip. You write a prompt, or you attach a still, and you get moving picture out. The catalog price is 2 credits per second of output, so a 5-second draft is about 10 credits. Durations are every whole second from 1s through 20s. Max output is 1080p. The published audio field is unset — do not call this native audio. That is the whole row. Confusing it with a Seedance pass (from 34 credits), or with a flat 2-credit remover, is how a cheapest-video search returns the wrong tool.

Two credits a second is a generate

Versely tags every row with a content_type. video means the output is a moving picture, not that the row made a clip from a prompt. Inside that pool sit generators and utilities.

P-Video is a generator: categories are text-to-video and image-to-video on one slug. requires_image is false, so the text path runs without a file. When the job is a still you already approved, attach it through image-to-video. When the job is a sentence, open the AI video generator. Same model, two contracts. Pick one per generate. A leftover PNG "for style" on the text path becomes a first-frame argument you did not mean to make.

The cheapest-video ranking filters to generators so a remover cannot win. That filter, and why it exists, is the cheapest video is not a one-credit remover. The live table is cheapest AI video model. Quote the table, not this sentence, if you need today's floor. In this snapshot the floor is this row’s 2-credit minimum job: one second of output.

Draft mode is how the spend stays a test: the description's speed claim is a 5s 720p clip in about 10 seconds. Qualities are 720p and 1080p. Aspects include 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, and 1:1. Frame rates are 24fps and 48fps. Draft short. Promote to 1080p and a longer take only after the motion is the motion you wanted.

Twenty seconds is the ceiling

Most published video rows cap at 15s. P-Video does not. The menu is 1s, 2s, 3s, then every second through 20s. That granularity is the point: you can buy 7s or 14s or 20s of output instead of rounding up to a 15s sibling and trimming.

Twenty seconds is the ceiling, not the default. A 20s generate at 2 credits a second is still a 20s mistake if the action is three ads glued in a prompt. Write one continuous motion, same as you would on a dearer long row. 1s is legal here. Use it to see whether the still holds or the sentence reads. Then pick 5s. Then, maybe, 20s.

Length is not quality. P-Video is the floor of a complete generate in the current published set. It is fast iteration with a 20s menu. It is not the row you pick because you need Seedance's 30-second identity pass, Veo's 4K eight seconds, or a guaranteed native mix.

If you needed native dialogue as a published field, this is not that row. Features mention audio-to-video and lip_sync; the audio field is still unset. Listen to the file, or pick a native-audio slug when speech is the product.

Not Seedance 2.5

Seedance 2.5 is 4–30s, 720p, from 34 credits (currently discounted from 67), native audio on. It is a single-pass generate that reasons about the whole shot so motion, lighting, and identity hold for thirty seconds. That is a different job and a different bill.

Do not treat P-Video's 20s as a discount Seedance. You will not get the same coherence, the same mix, or the extra ten seconds. You will get a long take at a low per-second rate, at 1080p. If the brief is "one coherent 30-second shot with native audio," you are on Seedance 2.5. If the brief is "I need seconds to see whether this motion works, at the lowest complete-generate cost we publish," you are on P-Video.

Grok Imagine Video is the other 30s menu (6 / 10 / 15 / 20 / 30s, 720p, from 24 credits, native audio). Still not this row. P-Video stops at 20s. Seedance and Grok are how you pass twenty. P-Video is how you stay cheap up to twenty.

Resolution is not a tie-breaker in Seedance's favor either. P-Video's ceiling is 1080p. Seedance 2.5's published ceiling is 720p. 4K is a different row on both families. Do not type 4K into a P-Video brief and expect a Veo plate.

Not a background remover

BRIO is content_type video, category edit-video, headline 2 credits. It removes a background from footage you already have. ByteDance's video upscaler starts at 3 credits, also a utility. Neither one answers "what is the cheapest model that will generate a clip."

Without a generator filter, sort-on-headline-price over content_type video crowns a remover. That is not the question. isVideoGenerator keeps rows that create video from text, image, or reference. P-Video passes. BRIO fails before sort. No amount of being cheap moves a remover onto a generate table.

If you actually need the background gone, open BRIO and pay that row. If you need a clip from a prompt or a still, you are on a generator. The cheapest complete job in this snapshot is P-Video: 2 credits per second of output, 1080p, up to 20 seconds. It will move when the catalog moves. The ranking page computes it.

Do not use cheapest-video as a quality pick. Cheap complete jobs can be short, silent, or low-fidelity. This page is only the correction: two credits a second is a real generator rate, and the 2-credit floor is a one-second job, not a twenty-second one. It is not Seedance. It is not a remover.

FAQ

Is P-Video really 2 credits for 20 seconds?

No. The published catalog price is 2 credits per second of output, so cost scales with clip length. Durations go to 20s at 1080p. Quote the live model page and the cheapest-video table if the snapshot has moved.

Do I have to upload an image?

No. requires_image is false. The row runs text-to-video without a file. If the job is "animate this still," upload it — otherwise you are not on the image-to-video contract.

Why isn't a background remover on the cheapest-video table?

Because it is not a generate. BRIO operates on footage you brought. Cheapest-ai-video-model filters to generators. A flat-priced utility cannot win a generate ranking.

Should I use this instead of Seedance 2.5?

Use it instead of Seedance when you need a low-rate 20s generate, not a 30s native-audio identity pass from 34 credits. Length, mix, and ceiling are different. Pick the row the job actually needs.