Skip to main content

​
About the Provider

Pruna AI is an AI optimization company focused on making state-of-the-art generative AI fast, affordable, and accessible. Their P-Video model is designed to deliver cinematic-quality video generation with configurable resolution, duration, FPS, and support for text, image, and audio-conditioned workflows.

​
Model Quickstart

This section helps you quickly get started with the p-video model on the Qubrid AI inferencing platform. To use this model, you need:
  • A valid Qubrid API key
  • Access to the Qubrid inference API
  • Basic knowledge of making API requests in your preferred language
Once authenticated with your API key, you can send inference requests to the p-video model and receive responses based on your input prompts. Below are example placeholders showing how the model can be accessed using different programming environments.
You can choose the one that best fits your workflow. 
import requests

url = "https://platform.qubrid.com/v1/videos/generations"
headers = {"Authorization": "Bearer QUBRID_API_KEY"}

# ─────────────────────────────────────────
# Case 1: JSON body — image URL + audio URL
# Works with text-only, image-only, audio-only, or both
# ─────────────────────────────────────────
response = requests.post(
    url,
    headers={**headers, "Content-Type": "application/json"},
    json={
        "model": "p-video",
        "prompt": "A butterfly flying through a flower garden",
        "duration": 5,
        "resolution": "720p",
        "fps": 24,
        "aspect_ratio": "16:9",
        "image": "https://example.com/input-image.jpg",   # optional
        "audio": "https://example.com/audio.mp3",         # optional
        "draft": False,
        "save_audio": True,
        "prompt_upsampling": True,
    },
)
print(response.json())

# ─────────────────────────────────────────
# Case 2: File upload — image_file + audio_file (multipart)
# Remove either file entry if not needed
# ─────────────────────────────────────────
with open("/path/to/your/image.jpg", "rb") as img, \
     open("/path/to/your/audio.mp3", "rb") as aud:
    response = requests.post(
        url,
        headers=headers,
        data={
            "model": "p-video",
            "prompt": "A butterfly flying through a flower garden",
            "duration": "5",
            "resolution": "720p",
            "fps": "24",
            "draft": "false",
            "save_audio": "true",
            "prompt_upsampling": "true",
        },
        files={
            "image_file": img,   # remove if not needed
            "audio_file": aud,   # remove if not needed
        },
    )
print(response.json())

# ─────────────────────────────────────────
# Tip: pass image/audio as URL via form fields (no file needed)
# files={} is required to force multipart/form-data
# ─────────────────────────────────────────
response = requests.post(
    url,
    headers=headers,
    data={
        "model": "p-video",
        "prompt": "A butterfly flying through a flower garden",
        "image": "https://example.com/input-image.jpg",   # optional
        "audio": "https://example.com/audio.mp3",         # optional
        "resolution": "720p",
        "fps": "24",
        "draft": "false",
        "save_audio": "true",
        "prompt_upsampling": "true",
    },
    files={},  # forces multipart/form-data — do not remove
)
print(response.json())

​
Example Generations

Below are some example videos generated by the P-Video model, showcasing its cinematic quality and capabilities:
Prompt: Generate a breaking news style video with the provided reporter face and audio with facial movements.
Prompt: Create a warm, emotional animated storytelling video using the provided image of a father sitting beside a child in bed reading a book at night. The father is gently reading aloud from the book, and his mouth should move naturally in sync with the narration audio. Story: The old clockmaker, Elias, lived in a world of ticking seconds, but his own time was running out. His shop smelled of brass and aged oil, filled with hundreds of clocks that hummed in unison. One rainy afternoon, a young girl entered, clutching a pocket watch that had stopped. “Can you fix it?” she asked. “It belonged to my grandfather.” Elias examined the watch. It wasn’t broken, just neglected. “It needs care, not repair,” he whispered. As he worked, he told her stories of the watches, how they held moments of joy and sorrow, beating only when someone remembered them. When he returned the watch, it was ticking steadily. The girl smiled, her eyes reflecting the shop’s light. Elias didn’t ask for payment. He knew the watch would keep ticking, and in that, he found his own timelessness.
Prompt: Transform the scene into a suspenseful alien attack sequence with eerie lighting shifts, shadows, atmospheric effects, character fear reactions synced to audio, dynamic camera motion, and cinematic VFX like distant alien silhouettes or lights.
Prompt: Create a 5-second cinematic ad of a high-performance GPU where its power is shown visually and explained through sound design dramatic lighting, slow rotating close-ups, spinning triple-fan cooling with a smooth, powerful airflow sound, subtle heat dissipation effects, fast digital whooshes and pulsing electronic tones representing high speed and AI processing, deep bass hits to emphasize power, clean mechanical hum for efficiency, smooth camera motion, reflective surfaces, bold tech-style color grading no text or UI overlays, features communicated through visuals and audio only.
Prompt: Create a cute animated video from the provided image where the child explains gravity with lip-synced audio, showing an apple falling from above due to gravity and gently bonking his head with a funny reaction, using bright lighting, smooth motion, and a playful cartoon style.
Prompt: Create a professional talking-head video using the provided female face image and audio, with accurate lip-sync, natural facial expressions, subtle head and eye movement, and place her in a clean, well-lit modern studio or cozy indoor setting with soft background blur, cinematic lighting, and smooth camera motion so it feels like a high-quality presenter explaining the audio.

​
Model Overview

P-Video is a premium AI video generation model developed by Pruna AI, designed for cinematic-quality video outputs.
  • It supports text-to-video, image-to-video, and audio-conditioned workflows with configurable resolution, duration, and FPS.
  • Prompt upsampling automatically enhances prompts for improved visual coherence and motion realism.
  • P-Video is suitable for production workflows requiring fast, scalable, and high-quality video generation across diverse creative and commercial use cases.

​
Model at a Glance

FeatureDetails
Model IDp-video
ProviderPruna AI
Model TypeVideo Generation
ArchitectureProprietary multi-modal video diffusion architecture
Release Date2026
LicenseProprietary
Output TypeVideo generated from text, image, or audio input

​
When to use?

You should consider using P-Video if:
  • You need cinematic-quality text-to-video generation
  • Your application requires image-to-video animation
  • You want audio-conditioned video storytelling
  • You are creating marketing, advertising, or social media content
  • You need rapid video prototyping with configurable quality settings

​
Inference Parameters

Parameter NameTypeDefaultDescription
Durationnumber5Duration of the video in seconds (1–10). Ignored when audio is provided.
Resolutionselect720pVideo resolution. Options: 720p, 1080p.
FPSselect24Frames per second. Options: 24 or 48.
Aspect Ratioselect16:9Aspect ratio of the video. Options: 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 1:1. Ignored when an input image is provided.
Seednumber0Random seed for reproducible video generation.
Draft ModebooleanfalseGenerate a lower-quality preview version of the video.
Save AudiobooleantrueWhether to include audio in the final generated video.
Prompt UpsamplingbooleantrueUse automatic prompt enhancement for better results.

​
Key Features

  • Text-to-Video Generation: Generates cinematic-quality videos from text prompts with configurable resolution, FPS, and duration.
  • Image-to-Video Animation: Accepts an input image to animate into a video sequence with motion synthesis.
  • Audio-Conditioned Generation: Supports audio input to condition video generation for synchronized storytelling.
  • Prompt Upsampling: Automatically enhances prompts for improved visual coherence and motion realism.
  • Flexible Input Support: Accepts JSON body with URLs or multipart file uploads for image and audio inputs.
  • Draft Mode: Generates lower-quality preview videos at reduced cost for rapid prototyping.

​
Limitations

  • Maximum duration limited to 10 seconds.
  • Audio formats limited to flac, mp3, and wav.
  • Higher resolution increases generation time.
  • Sync mode may timeout for longer videos.

​
Summary

P-Video is a premium AI video generation model built by Pruna AI for cinematic-quality video outputs.
  • It supports text-to-video, image-to-video, and audio-conditioned workflows with configurable resolution up to 1080p.
  • The model offers flexible input support via JSON body or multipart file uploads.
  • It is suitable for developers and enterprises requiring scalable, high-quality video generation for creative and commercial workflows.