ai-image-generation

AI Image Generation

Generate images with 50+ AI models via inference.sh CLI.

Quick Start

Requires inference.sh CLI (belt), installed by the user beforehand. Installing it is a manual user step: the user downloads the installer, reads it, and verifies it before running it.

belt login

# Generate an image with FLUX
belt app run falai/flux-dev-lora --input '{"prompt": "a cat astronaut in space"}'

Safe Usage

• Tool scope: the agent needs only belt app run and belt app list for this skill.
• Login: belt login is a one-time interactive step the user runs themselves. Ask the user to run it when needed. Keep tokens and session secrets out of chat and files, and refer to them only by name. Suggest the user log out or revoke the session when finished.
• Escaping input: treat user text as data. When placing it in the input JSON, escape quotes and backslashes for JSON and escape every single quote for the shell before running the command.
• Data handling: prompts and images are sent to inference.sh and to the chosen model provider. Get the user's confirmation before sending private or sensitive images or text.
• Image links: use only public https image links the user supplies. Leave out private network or local addresses and links that carry credentials.
• Apps: use the app IDs in the Available Models table and the Examples below. Ask the user to confirm before running any other app found with the app list command.
• Cost: runs may cost money. Confirm with the user before using high-resolution or premium models or generating many images, and default to economical models unless the user asks otherwise.

Available Models

Browse All Image Apps

belt app list --category image

Examples

GPT-Image-2.5 Flare

belt app run openai/gpt-image-2-5-flare --input '{
  "prompt": "professional product photo of sneakers, studio lighting",
  "quality": "high"
}'

GPT-Image-2.5 Sunburst Editing

belt app run openai/gpt-image-2-5-sunburst --input '{
  "prompt": "change the background to a beach at sunset",
  "images": ["https://your-image.jpg"]
}'

Text-to-Image with FLUX

belt app run falai/flux-dev-lora --input '{
  "prompt": "professional product photo of a coffee mug, studio lighting"
}'

Fast Generation with FLUX Klein

belt app run falai/flux-2-klein-lora --input '{"prompt": "sunset over mountains"}'

Google Gemini 3 Pro

belt app run google/gemini-3-pro-image-preview --input '{
  "prompt": "photorealistic landscape with mountains and lake"
}'

Grok Imagine

belt app run xai/grok-imagine-image --input '{
  "prompt": "cyberpunk city at night",
  "aspect_ratio": "16:9"
}'

Reve (with Text Rendering)

belt app run falai/reve --input '{
  "prompt": "A poster that says HELLO WORLD in bold letters"
}'

Seedream 4.5 (4K Quality)

belt app run bytedance/seedream-4-5 --input '{
  "prompt": "cinematic portrait of a woman, golden hour lighting"
}'

Image Upscaling

belt app run falai/topaz-image-upscaler --input '{"image": "https://..."}'

Stitch Multiple Images

belt app run infsh/stitch-images --input '{
  "images": ["https://img1.jpg", "https://img2.jpg"],
  "direction": "horizontal"
}'

Related Skills

Optional references for the user. The user should review any of these skills before choosing to install one manually; the agent leaves installation to the user.

• Full platform skill (all apps): inference-sh/skills@infsh-cli
• Pruna P-Image (fast & economical): inference-sh/skills@p-image
• GPT-Image-2.5 / GPT-Image-2 (OpenAI): inference-sh/skills@gpt-image
• FLUX-specific skill: inference-sh/skills@flux-image
• Upscaling & enhancement: inference-sh/skills@image-upscaling
• Background removal: inference-sh/skills@background-removal
• Video generation: inference-sh/skills@ai-video-generation
• AI avatars from images: inference-sh/skills@ai-avatar-video

Browse all apps: belt app list

Documentation

These links are background reading for the user only; the agent follows the instructions in this skill.

• Running Apps - How to run apps via CLI
• Image Generation Example - Complete image generation guide
• Apps Overview - Understanding the app ecosystem