fal-image-edit（fal 图像编辑）

fal.ai Image Edit

Edit images using AI: style transfer, object removal, background changes, and more.

How It Works

1. User provides image URL and editing instructions
2. Script selects appropriate model
3. Sends request to fal.ai API
4. Returns edited image URL

Finding Models

To discover the best and latest image editing models, use the search API:

# Search for image editing models
bash /mnt/skills/user/fal-generate/scripts/search-models.sh --category "image-to-image"

# Search for specific editing capabilities
bash /mnt/skills/user/fal-generate/scripts/search-models.sh --query "image editing"
bash /mnt/skills/user/fal-generate/scripts/search-models.sh --query "inpainting"
bash /mnt/skills/user/fal-generate/scripts/search-models.sh --query "object removal"
bash /mnt/skills/user/fal-generate/scripts/search-models.sh --query "background removal"

Or use the search_models MCP tool with relevant keywords.

Supported Operations

Usage

bash /mnt/skills/user/fal-image-edit/scripts/edit-image.sh [options]

Arguments:

• --image-url - URL of image to edit (required)
• --prompt - Description of desired edit (required)
• --operation - Edit operation: style, remove, background, inpaint (default: style)
• --mask-url - URL of mask image (required for inpainting/removal)
• --strength - Edit strength 0.0-1.0 (default: 0.75)

Examples:

# Style transfer
bash /mnt/skills/user/fal-image-edit/scripts/edit-image.sh \
  --image-url "https://example.com/photo.jpg" \
  --prompt "Convert to anime style" \
  --operation style

# Remove object
bash /mnt/skills/user/fal-image-edit/scripts/edit-image.sh \
  --image-url "https://example.com/photo.jpg" \
  --prompt "Remove the person on the left" \
  --operation remove

# Change background
bash /mnt/skills/user/fal-image-edit/scripts/edit-image.sh \
  --image-url "https://example.com/portrait.jpg" \
  --prompt "Place in a tropical beach setting" \
  --operation background

# Inpainting with mask
bash /mnt/skills/user/fal-image-edit/scripts/edit-image.sh \
  --image-url "https://example.com/photo.jpg" \
  --mask-url "https://example.com/mask.png" \
  --prompt "Fill with flowers" \
  --operation inpaint

MCP Tool Alternative

Use search_models MCP tool or search-models.sh to find the best current model for each operation (style transfer, object removal, background change, inpainting), then call mcp__fal-ai__generate with the discovered modelId.

Output

Editing image...
Model: fal-ai/flux/dev/image-to-image
Operation: style transfer

Edit complete!

Image URL: https://v3.fal.media/files/abc123/edited.png
Dimensions: 1024x1024

JSON output:

{
  "images": [
    {
      "url": "https://v3.fal.media/files/abc123/edited.png",
      "width": 1024,
      "height": 1024
    }
  ]
}

Present Results to User

Here's your edited image:

![Edited Image](https://v3.fal.media/files/...)

• 1024×1024 | Operation: Style Transfer

Model Selection Tips

• General Editing: Search for "image editing" models. Good for instruction-based changes.
• Style Transfer: Search for image-to-image category. Adjust strength: 0.3-0.5 for subtle, 0.7-0.9 for dramatic.
• Object Removal: Search for "eraser" or "object removal". Some work without masks.
• Background Change: Search for "background" or "kontext". Look for models that preserve subject identity.
• Inpainting: Search for "inpainting" or "fill". Requires binary mask (white = edit area).

Mask Tips

For inpainting and some removal tasks:

• White pixels = areas to edit
• Black pixels = areas to preserve
• Use PNG format with transparency or solid colors
• Feathered edges create smoother transitions

Troubleshooting

Edit Too Subtle

The edit is barely visible.

Increase the strength parameter:
--strength 0.85

Edit Too Dramatic

The edit changed too much of the image.

Decrease the strength parameter:
--strength 0.3

Object Not Removed

The object wasn't fully removed.

Tips:
1. Be more specific in the prompt
2. Try using an explicit mask
3. Use the inpainting model for precise control

Background Artifacts

The new background has artifacts around the subject.

Tips:
1. Use a cleaner source image
2. Try FLUX Kontext which handles edges better
3. Adjust the strength for smoother blending