--- title: "P-Video" description: "P-Video on Kuro: Pruna's general clip model. Cheap, flexible durations, one reference audio. From 8.5 credits for five seconds (≈ $0.20 on Starter)." canonical_url: "https://meetkuro.com/models/p-video/" locale: "en" model: "p-video" --- # P-Video Pruna AI · Shorts · standard Pruna's general clip model. Cheap, flexible durations, one reference audio. ## What it costs - 5s · 720p: 8.5 credits (≈ $0.20) - 5s · 1080p: 17 credits (≈ $0.40) Dollar figures are approximate, at the Starter plan's rate of $19.00 for 800 credits. ## What it takes - **Aspect ratios**: 16:9 9:16 4:3 3:4 3:2 2:3 1:1 - **Resolutions**: 720p 1080p - **Duration**: 1–10 seconds - **Reference images**: Not accepted - **Reference audio**: 1 - **Last frame**: Accepted ## When to reach for it The economy option for general clip work: seven aspect ratios, one to ten seconds, 720p or 1080p, and a last-frame slot for chaining. At a fraction of what the premium models charge, it is a sensible place to work out timing and motion before spending real credits. It takes no reference images, which is the limitation to plan around: a subject cannot be carried into it. It does accept one reference audio, which is unusual at this price and means a clip can be cut to a track or a voice line you already have. The wide duration range is its most useful property. One second is a real option, and short punctuating cuts are exactly the kind of shot where paying premium rates makes least sense. Use it for the connective tissue of a sequence and spend the budget on the shots people will actually look at. If it keeps missing, the sibling models (Avatar, Animate and Replace) are purpose-built for specific jobs and will beat it at those. ## How to prompt P-Video Describe what you want in your own words and Kuro writes the prompt for you. If you would rather write it yourself, this is what P-Video responds to. P-Video works from text alone, animates a first frame, travels from a first frame to a last one, or follows an attached audio track. Start from the provider's own formula — Subject, Action, Scene, Camera, Lighting, Style, Audio — of which only the first three are required. Lead with one clear subject action, then the camera move and the sound. Name what is visible rather than the camera term for it, and describe anyone who must stay the same once, in detail, then point back to that. Keep each take to one or two beats of close, foreground action; it is strongest on close-up subjects and objects in the foreground. One continuous shot per take — several cuts scripted into one prompt come back muddled. With a first frame, write only what changes; a last frame gives the motion a destination. If the shot is silent, say so, and say the mouth stays closed: a moving mouth over a silent track reads as dubbed. An attached audio track decides the clip's length, which suits dialogue, performance timing and music-driven motion — and is a cheap way to cut a clip to music. Draft at 720p; dropping the tier changes sharpness and nothing the prompt controls. Vertical work is reported to hold up better at 1080p if a 9:16 take looks soft. The same advice covers [P-Video 2](https://meetkuro.com/models/p-video-2.md). Their pages note what differs. ### What it can't do - No reference images or videos. For consistency across takes, settle the subject in one still and use it as the first frame. - One audio track at most. Sound effects are the weaker half of its sound; where the sound is the product, make the track separately and attach it. - The provider documents its own limits: no extreme cinematic camera motion, and speaker attribution that drifts above two voices. Do not plan a multi-scene story in one take. - Chaining takes by feeding each clip's last frame into the next makes the face drift and lose colour with every hop. Build each part from its own fresh still, and never chain frames for speech. - A gesture in the starting still plays as an opening beat and returns at random, not at a moment you choose. - No negative prompt. ### Use something else when… - A portrait should talk over a voice track: [P-Video Avatar](https://meetkuro.com/models/p-video-avatar.md). - A character should perform motion from a video you already have: [P-Video Animate](https://meetkuro.com/models/p-video-animate.md). - The same shot should have a different person in it: [P-Video Replace](https://meetkuro.com/models/p-video-replace.md). - You need a character held by reference images, extreme camera moves, or a long multi-beat take: [Seedance 2.5](https://meetkuro.com/models/seedance-2.5.md). ## Guides that cover it - [Two pictures, fifteen seconds: the first-and-last-frame method](https://meetkuro.com/guides/first-and-last-frame.md): Make the opening frame and the closing frame yourself, hand a video model both, and let it fill the middle. Most models take a last frame. What it buys. - [P-Video: the complete guide to Pruna's four models](https://meetkuro.com/guides/p-video-prompting.md): Four models, one family: a cheap generator and three performance rigs. What each one takes, what a take really costs, and the failures worth knowing first. ## P-Video is one pick in the composer. Every model in this catalog runs from the same prompt bar, priced in the same credits. --- ## About Kuro Kuro is an AI picture studio: prompt-driven images, clips, voice-overs and music, and storyboard-backed films with consistent characters across every scene. Agents connect over MCP and can generate directly: ``` claude mcp add kuro https://api.meetkuro.com/mcp --transport http ``` The app is at https://app.meetkuro.com and requires an account. Each guide, model page, comparison and free tool page on this site (not a tool's variants, and not the index pages) has a Markdown twin at its own URL plus `.md`, and so do /benchmarks/, the English /agents/ page (as /agents.md) and the MCP tool reference. Model catalog: https://meetkuro.com/models/