Pruna P-Video: 2 credits per second, 1–20s

PrunaAI P-Video is 2 credits per second of output, for clips of 1–20 seconds at 1080p. Audio is unset. A draft row, not a native-audio hero. Use it to see whether the motion exists. Do not use it as a stand-in for Veo, Seedance, or Kling O3 Pro.

PrunaAI P-Video is content type video, categories text-to-video and image-to-video, image not required, durations every whole second from 1s through 20s, max output 1080p, 2 credits per second of output, audio null. Description also mentions draft mode, a 5s 720p clip in about 10 seconds, and audio-to-video in a single endpoint. The published audio field is still unset. Do not call this native audio. Listen to the file if a track appears. Do not brief a monologue as if the catalog promised one.

Twenty seconds is the menu. Two credits a second is the sticker. Neither is a reason to skip the rows that actually write speech.

Draft length is the product

Most published video rows cap at 10s or 15s, or they sell 6s and 10s as stairs. P-Video sells 7s. And 14s. And 20s. At 2 credits a second. That granularity is for drafts: one continuous motion, long enough to see whether the idea holds, cheap enough to throw away at the short end.

Use 1s–5s to see if a still holds or a sentence reads. The description’s speed claim is a 5s 720p clip in about 10 seconds. Draft short.

Use 6s–12s when the motion is one action (orbit, pour, walk) and you are not ready to spend on a hero native-audio pass (Seedance 2.5 from 34 credits, Seedance 2.0 from 44).

Use 20s when the action is actually 20 seconds. Three ads glued in a prompt is still three ads. Twenty seconds of a drifting face is a cheap long mistake.

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. A leftover PNG “for style” on the text path becomes a first-frame argument you did not mean to make.

1080p is the ceiling. Qualities in the family include 720p and 1080p. You are not on a 4K hero. If the delivery sheet says 4K, either generate on a 4K row or upscale after lock. ByteDance Upscaler Video starts at 3 credits and targets up to 4K. Do not upscale a 20s draft you will throw away.

This is a generate. You did not already have the clip. Restyle, removers, and upscalers are edits that start at 2 to 4 credits. P-Video is a generate at 2 credits a second because it invents motion. Do not put it on a cheapest-edit table.

Audio is unset on purpose

Native audio on Versely is a catalog flag: the generate wrote a track. P-Video’s flag is null. Features may mention audio-to-video and lip_sync. The field you route on is still unset.

Practical rules:

• Do not write dialogue into the prompt and wait for a mix.
• Do not skip a TTS plus lipsync plan because “it says audio-to-video in the description.”
• If a file comes back with sound, treat it as a bonus you still have to QC, not as a contract.
• If we see a mouth, budget a lipsync row: VEED Lipsync from 3 credits on a take, or Fabric / Avatar Pro on a still.

Hero native-audio rows, when the line is the product:

• Veo 3.1 — from 64 credits, 4K, 4/6/8s, native audio.
• Seedance 2.0 — from 44 credits, 1080p, 4–15s, native audio.
• Kling O3 Pro text-to-video — from 56 credits, 3–15s, native audio.
• Seedance 2.5 — from 34 credits, 720p, 4–30s, native audio, if you needed thirty seconds with a track.

A low per-second rate versus a hero row is not a saving if the brief was a spoken hook. You will pay the TTS and the mouth anyway, or you will ship a silent 20s file into a mute feed and wonder why no one stayed.

Silent autoplay still wins even when a hero did write audio. Captions for silent autoplay. On a draft row, captions are how you test the line without a mix. Add captions to video. Caption timing on a 20s draft will tell you the sentence is too long before you spend on a hero pass.

Do not confuse 20s with a 30s hero

Seedance 2.5 is the 30-second native-audio single pass from 34 credits, 720p. P-Video is 20s, 1080p, 2 credits a second, audio null. Length is not the same job as identity-plus-speech for half a minute.

Hailuo 2.3 Pro is 6s and 10s, 1080p, from 40 credits, audio false. Runway Gen-4.5 is 5s and 10s, from 200 credits, audio false. Those are silent too, and they are not 20s, and they are not 2 credits a second. P-Video is the low-rate long silent draft. They are other silent rows with other reasons (control, 1080p “pro” label). Pick the reason.

A 20s P-Video file on a 4K timeline next to Veo 8s native-audio plates will look like what it is: a draft. Either finish it (upscale, VO, lipsync, captions) or replace it. Shipping the draft because it was 2 credits is how a week of tests becomes the campaign.

UGC-shaped tests are the honest use. Kapwing’s 2026 UGC figures still put human authenticity above AI (81% vs 63% in that set). An 8-second motion test (about 16 credits at 2 a second) under a real VO is closer to a creator workflow than a talking generate you have not timed. Cite the stat, then caption. Do not claim P-Video is UGC.

Photoreal drafts of people: EU Article 50 from 2 August 2026 if you publish. Article 50 disclosure. Draft is not unpublished if it is on the feed.

Confirm 20s only after 5s works

Open the catalog on PrunaAI P-Video. Confirm 2 credits per second, 1080p, a duration between 1s and 20s, audio unset. Run 5s. If the motion is the motion, run the length you will ship, still on this row, or promote to a native-audio hero if speech is now the brief.

If the confirm screen is a flat-priced remover or restyle, you drifted into an edit. If it is a Seedance row with native audio on, you are on a hero — use it when the draft earned it, not as a first click.

FAQ

Does P-Video write native audio?

The published audio field is null. Do not treat it as a native-audio hero. Veo, Seedance, and Kling O3 Pro are the catalog flags for that job.

Why would I pick P-Video instead of LTX 2.3 Fast?

LTX 2.3 Text to Video Fast starts at 20 credits (6–20s clips, 4 credits a second at SD/HD and 13 at 4K, native audio true). P-Video is 2 credits a second, 1–20s, 1080p, audio unset, every whole second including 1s–5s. Short drafts and a lower per-second rate are the P-Video case. A 4K pass with audio is Fast.

Is 20 seconds the default?

No. It is the ceiling. Start at 1s or 5s. Buy 20s when the action is 20s. The low per-second rate does not make a 20s prompt with three scenes a good idea.