P-Image
Pruna AI · AI imageP-Image is Pruna AI's premium image generation and editing model, delivering state-of-the-art results in less than one second. It offers a rare combination of speed, affordability, visual quality, exact prompt adherence, and exceptional text rendering.
P-Image intelligently switches between two Replicate endpoints: pure text-to-image generation when no reference image is provided, and image-with-text editing when a reference image is supplied — all behind a single model ID at a flat rate of 2 credits per image.
Key features and benefits
Dual-mode generation
Use P-Image for text-to-image generation with just a prompt and preferred aspect ratio. Add a reference image and it automatically switches to image editing mode — describe your edit in the prompt and P-Image-Edit handles the rest.
Blazing fast
P-Image is tuned for sub-second inference, making it ideal for high-volume image workflows where turnaround time matters.
Text rendering
Exceptional prompt adherence and accurate text rendering built in — logos, signs, and on-screen text come out crisp and readable.
Aspect ratio control
Choose from standard aspect ratios (1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3). In editing mode, aspect_ratio defaults to matching the input image.
Technical specifications
ProviderPruna AI (P-Image)
Primary inputsText prompt (required); reference image(s) optional for editing mode
Supported aspect ratios1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, custom; edit mode supports match_input_image
Max resolution1440×1440 (custom dimensions available via custom aspect ratio)
Use cases and applications
Generate scene start frames for AI Video projects at a flat 2 credits per image, whether you're creating from a prompt or editing an existing frame.
Edit production frames by supplying a reference image and describing the change — P-Image-Edit handles single-image editing with high precision.
Why this model
Choose P-Image when you want the fastest possible image generation and editing at a predictable 2-credit flat rate, with strong prompt adherence and text rendering.
Its dual-mode design means you don't need separate models for generation and editing — just pass (or omit) a reference image and the model adapts automatically.
How VidMachine uses it
VidMachine uses P-Image for text-to-image scene start frames when no reference image is available, and seamlessly switches to P-Image-Edit when a reference image (e.g. a project style reference) is provided.
Generation runs on Replicate at 2 credits per image, flat rate regardless of mode.
What you should know
What's the difference between P-Image and P-Image-Edit?
They're the same model ID in VidMachine. P-Image uses Replicate's text-to-image endpoint (prunaai/p-image) when no reference image is given, and the image-edit endpoint (prunaai/p-image-edit) when one or more reference images are provided. Billing is always 2 credits.
How many reference images can I use in edit mode?
P-Image accepts up to 3 reference images in edit mode. Provide the main image to edit as the first image, and describe your edit clearly in the prompt.
Can I use custom dimensions?
Yes. Select aspect_ratio custom and provide width and height (both must be multiples of 16, up to 1440). Videos in 9:16 or 16:9 are the most common VidMachine use cases.
How is P-Image billed?
2 credits per generated image, flat rate whether you use text-to-image or image-edit mode.