Pruna AI family
# Pruna Video 2 High-fidelity generative video for professional marketing. ## Overview Pruna Video 2 is a state-of-the-art generative video model designed to transform text, images, and audio into high-definition cinematic content. As the successor to the original P-Video, this model focuses on superior visual quality, temporal consistency, and high-frame-rate output, supporting up to 1080p resolution at 48 frames per second. It is engineered to meet the demanding requirements of ad agencies and creative teams who need reliable, high-quality video assets produced at scale. Beyond simple text-to-video generation, Pruna Video 2 offers advanced conditioning capabilities, including image-to-video synthesis and native audio integration. The model excels at maintaining structural integrity across frames, making it ideal for complex motion sequences. With features like prompt upsampling and optional lip-syncing, users can achieve professional-grade results with minimal manual intervention. Whether you are creating short-form social media ads or high-resolution product showcases, Pruna Video 2 provides the flexibility and control necessary for modern digital storytelling. ## Use Cases * **Social Media Ad Campaigns:** Generate eye-catching, high-frame-rate video ads for platforms like TikTok and Instagram Reels from simple text prompts. * **Product Visualization:** Convert static product photography into dynamic, 1080p motion sequences to increase engagement on e-commerce landing pages. * **Automated Lip-Sync Content:** Create personalized video messages or influencer-style content by syncing generated video motion to provided audio tracks. * **Rapid Prototyping:** Use the draft mode to quickly iterate on video concepts and visual styles before committing to high-resolution, full-length renders. * **Brand Storytelling:** Animate brand assets or static illustrations to create cohesive, professional-looking video sequences for digital marketing campaigns. * **Content Localization:** Generate consistent visual backgrounds or B-roll footage that can be easily adapted for different regional marketing campaigns. ## Parameters Pruna Video 2 offers a comprehensive suite of input parameters to fine-tune your video output. Users can control technical specifications such as resolution (720p or 1080p), frame rate (24 or 48 fps), and aspect ratio. Creative control is managed through text prompts, image conditioning, and audio input, while technical parameters like seed, duration, and safety filter toggles allow for precise, reproducible, and safe generation workflows. Please refer to the Pixloop interface for the full list of available configuration fields. ## Tips for Best Results * **Leverage Prompt Upsampling:** Keep the 'prompt_upsampling' feature enabled to allow the model to expand on your core ideas, resulting in richer visual detail and better adherence to your creative intent. * **Use Draft Mode for Iteration:** When testing new concepts, use the 'draft' mode to generate low-cost previews. This saves time and resources while you refine your prompt or visual composition. * **Consistent Aspect Ratios:** Always match your 'aspect_ratio' setting to the target platform (e.g., 9:16 for mobile stories, 16:9 for YouTube) to avoid unnecessary cropping or letterboxing. * **Condition with Images:** For specific product shots or brand-consistent characters, provide an input image to anchor the generation process and ensure the model maintains the desired visual identity. * **Manage Duration Carefully:** If you are not providing an audio track, use the 'duration' parameter to define the exact length of your clip. For audio-driven videos, the model will automatically determine the length based on the input file. ## About Pruna Video 2 is developed by PrunaAI and is available via the Replicate infrastructure. It represents a significant advancement in quality-focused generative video, building upon the foundation of the original P-Video model. Designed for efficiency and high-performance output, it is optimized for professional creative workflows where visual fidelity and temporal stability are paramount.
Model input reference derived from preset schema.
| Name | Type | Required | Default | Description |
|---|---|---|---|---|
| fps | select | No | 24 | Frames per second of the video. |
| seed | integer | No | null | Random seed. Set for reproducible generation. |
| audio | audio | No | null | Input audio to condition video generation. Supports flac, mp3, wav. |
| draft | boolean | No | false | Draft mode. Generates a lower-quality preview of the video. |
| image | image | No | null | Input image to generate video from (image-to-video). Supports jpg, jpeg, png, webp. |
| no_op | boolean | No | false | Health check mode - returns status without inference. |
| prompt | string | Yes | null | Text prompt for video generation. |
| duration | integer | No | null | Duration of the video in seconds (1-20). Leave empty to let the model choose length from the prompt. Ignored when audio is provided. |
| resolution | select | No | "720p" | Resolution of the video. |
| save_audio | boolean | No | true | Save the video with audio. |
| aspect_ratio | select | No | "16:9" | Aspect ratio of the video. Ignored when an input image is provided. |
| last_frame_image | image | No | null | Reference image for the last frame of the video. Supports jpg, jpeg, png, webp. |
| prompt_upsampling | boolean | No | true | Use prompt upsampling to enhance the prompt. |
| disable_safety_filter | boolean | No | true | Disable safety filter for prompts (and input image). When disabled, prompts are not checked for unsafe content before generation. |