ImageFlare Skill

Generate and edit images from your terminal using Cloudflare Workers AI models (Flux, Stable Diffusion, etc.).

When to Use

✅ USE this skill when:

User asks to generate an image from a text description
User wants to edit or transform an existing image with AI
User wants to apply a style from a reference image to another image
User asks to configure Cloudflare AI credentials or switch models
User says "create an image", "generate a picture", "edit this photo with AI"

When NOT to Use

❌ DON'T use this skill when:

Local image manipulation without AI (crop, resize, rotate) → use convert/ffmpeg/Pillow
Video generation → not supported
Non-Cloudflare AI image generation → use other tools
Viewing or inspecting image metadata → use identify, exiftool
The user has not configured Cloudflare credentials yet and doesn't want to → prompt them to run imageflare config first

Setup

Requires a Cloudflare account with Workers AI access (free tier available).

# First-time interactive setup (Account ID + API Token + model selection)
imageflare config

Getting credentials:

Credential	Where to find it
Account ID	Cloudflare dashboard → right sidebar of any page
API Token	My Profile → API Tokens → Create Token — select the Workers AI template or grant `Workers AI: Read` permission

Verify setup:

imageflare config show

Commands

Generate an Image

# Basic generation
imageflare generate --prompt "a red fox sitting on a snow-covered log"

# With custom dimensions and seed for reproducibility
imageflare generate --prompt "a sunset over mountains" --width 512 --height 512 --seed 42

# Save to specific path and auto-open
imageflare generate --prompt "cyberpunk cityscape" --output cityscape.png --open

# Use a specific model
imageflare generate --prompt "a cat" --model "@cf/black-forest-labs/flux-1-schnell"

Edit an Existing Image

# Basic edit
imageflare edit photo.png --prompt "change the background to a beach"

# Edit with a style reference image (up to 3 refs)
imageflare edit photo.png --ref style.png --prompt "style image 0 like image 1"

# Multiple reference images
imageflare edit photo.png --ref ref1.png --ref ref2.png --prompt "combine styles"

# Save to specific path
imageflare edit photo.png --prompt "make it a watercolor painting" --output watercolor.png --open

Configuration

# Interactive setup wizard
imageflare config

# View current settings
imageflare config show

# List available AI models on your account
imageflare config models

# Set values non-interactively
imageflare config set --account-id YOUR_ID --api-token YOUR_TOKEN
imageflare config set --model "@cf/black-forest-labs/flux-1-schnell"

Common Options

Flag	Description
`-p, --prompt`	Text prompt describing what to generate or how to edit (required)
`-m, --model`	Cloudflare Workers AI model ID (overrides configured default)
`--width`	Output width in pixels (default: 1024)
`--height`	Output height in pixels (default: 1024)
`--seed`	Random seed for reproducible results
`-o, --output`	Output file path (default: `imageflare_<timestamp>.png`)
`--open`	Open the saved image automatically after generation
`-r, --ref`	(edit only) Additional reference image, repeatable up to 3 times

Output

Images are saved as PNG files
Default filename: imageflare_<timestamp>.png in the current directory
Use --output to specify a custom path
Use --open to auto-open the result in the system image viewer

Notes

Default model: @cf/black-forest-labs/flux-2-klein-4b
Input images for editing are automatically resized to ≤512×512 (Cloudflare requirement)
Reference images in edit mode: refer to them in your prompt as image 0, image 1, etc.
Config is stored at ~/.config/imageflare/config (Linux/macOS) or %APPDATA%imageflareconfig (Windows)
No intermediate servers — requests go directly to the Cloudflare Workers AI API
Free tier available on Cloudflare Workers AI