p-image
Generate images with Pruna P-Image models via inference.sh CLI. Models: P-Image, P-Image-LoRA, P-Image-Edit, P-Image-Edit-LoRA. Capabilities: text-to-image, image editing, LoRA styles, multi-image compositing, fast inference. Pruna optimizes models for speed without quality loss. Triggers: pruna, p-image, pruna image, fast image generation, optimized flux, pruna ai, p image, fast ai image, economic image generation, cheap image generation
Security Assessment
About p-image
The p-image skill generates and edits images using Pruna's optimized P-Image models through the inference.sh CLI (belt). Pruna optimizes AI models for speed without sacrificing quality, so the skill targets fast, economical text-to-image generation, image editing, LoRA-styled output, and multi-image compositing from the command line. The basic workflow is belt login followed by belt app run pruna/p-image with a JSON --input containing a prompt and options such as aspect_ratio.
Four P-Image models are documented: P-Image (pruna/p-image) for fast text-to-image with multiple aspect ratios; P-Image-LoRA (pruna/p-image-lora) for custom styles via preset LoRAs; P-Image-Edit (pruna/p-image-edit) for image editing with multi-image support; and P-Image-Edit-LoRA (pruna/p-image-edit-lora) for stylized editing. P-Image-LoRA includes built-in presets such as photos-realism, pixel-art, japanese-modern-look, cinematic-movie-style, graffiti-splash, neon-punk, anime-2-5d, ethereal-portrait, retro-90s-style, ink-sketchbook, and paper-cut. Editing accepts an images array (single image editing or multi-image compositing into a collage). Supported aspect ratios are 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, or custom with width/height. The doc also lists other optimized Pruna models runnable through belt, including flux-dev, flux-klein-4b (extremely fast at $0.0001/image), qwen-image, z-image-turbo, and wan-image-small, and users can browse Pruna apps with belt app list --namespace pruna.
The skill targets designers, marketers, and developers who want quick, low-cost image generation and editing. It requires the inference.sh CLI (belt), installable via npx skills add belt-sh/cli, and limits its allowed tool to Bash(belt *). Related skills cover the full inference.sh platform, all image-generation models, Pruna video generation, and FLUX models.
FAQ
Which P-Image models are available?
Four: P-Image (pruna/p-image) for fast text-to-image, P-Image-LoRA (pruna/p-image-lora) for preset LoRA styles, P-Image-Edit (pruna/p-image-edit) for editing with multi-image support, and P-Image-Edit-LoRA (pruna/p-image-edit-lora) for stylized editing.
What LoRA style presets are included?
P-Image-LoRA includes photos-realism, pixel-art, japanese-modern-look, cinematic-movie-style, graffiti-splash, neon-punk, anime-2-5d, ethereal-portrait, retro-90s-style, ink-sketchbook, and paper-cut.
What aspect ratios are supported?
1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, or custom with a specified width and height.
Can I edit or combine existing images?
Yes. P-Image-Edit accepts an images array — pass one image URL to edit it (for example changing the background) or several image URLs to composite them into a collage.
What do I need to run it?
The inference.sh CLI called belt, installed via npx skills add belt-sh/cli, then belt login. The allowed tool is Bash(belt *). Other optimized Pruna models such as flux-klein-4b ($0.0001/image), z-image-turbo, and qwen-image are also runnable through belt.
Install p-image
Quick Setup:
- Copy the skill folder to
.claude/skills/ - Claude will automatically detect and use the skill
Repository
101-skills/skills