Create an Image (SDXL)#

Generate high-quality images from text descriptions using the powerful SDXL model. This tool allows you to create detailed images from text prompts with support for custom models (LoRAs), style transfer, and shape guidance features.

Overview#

Text-to-image generation using SDXL
Custom model (LoRA) support
Style image transfer using ipAdapter
Shape guidance with ControlNet
Multiple aspect ratio options
Starting image modifications

Getting Started#

Visit Eden Create: Create an image (SDXL) tool
Enter your text prompt describing the image you want to create
Adjust settings like dimensions, guidance scale, and steps
Click "Create" and wait for your image to generate

Example Prompts#

"A majestic dragon soaring over a medieval castle, oil painting style"
"A futuristic cityscape at sunset with flying cars and neon lights"
"A serene forest clearing with morning mist and golden sunlight"

Using with Your Agents#

Ask your agent to "Create an image (SDXL)".

Basic Usage#

Required parameters:

prompt: Text description of desired image
width: Image width in pixels (256-2048)
height: Image height in pixels (256-2048)

Optional parameters:

n_samples: Number of images to generate (1-4)
steps: Number of diffusion steps (10-50)
seed: Random seed for reproducibility

Technical Guidelines#

Supported Resolutions#

Aspect Ratio	Dimensions	Use Case
21:9	1536x640	Cinematic
16:9	1344x768	Widescreen
3:2	1216x832	Standard
4:3	1152x896	Classic
1:1	1024x1024	Square
9:16	768x1344	Mobile
9:21	640x1536	Panoramic

Best Practices#

Prompting Tips#

Start with the main subject
Include specific details about style and composition
Add artistic medium and quality descriptors
Use clear, concise language

Resolution Guidelines#

Stay close to 1 megapixel for optimal quality
Use dimensions divisible by 8
Match aspect ratio to content type

Advanced Features#

Custom Models (LoRA)#

Enable custom model support with:

use_lora: Set to true
lora: Specify LoRA model name
lora_strength: Control influence (0.0-1.0)

Style Transfer#

Apply style from reference images:

use_ipadapter: Set to true
style_image: Provide reference image
ipadapter_strength: Adjust influence (0.0-1.5)

Shape Guidance#

Guide generation with input image shapes:

use_controlnet: Set to true
init_image: Provide starting image
controlnet_strength: Adjust influence (0.0-1.2)
preprocessor: Choose method (canny, depth, scribble, etc.)

Technical Details#

Base Configuration#

Model: SDXL
Default steps: 35
Default resolution: 1024x1024
Cost estimate: 1 * n_samples

Limitations#

Maximum resolution: 2048x2048
Maximum samples: 4
May struggle with text rendering
Complex prompts can be challenging

API Reference#

Endpoint#

curl -X POST "https://api.eden.art/v2/tasks/create" \
  -H "Content-Type: application/json" \
  -H "X-Api-Key: YOUR_API_KEY" \
  -d '{
    "tool": "txt2img",
    "args": {
      "prompt": "A captivating and vibrant illustration of an abstract face, predominantly focusing on the mesmerizing eyes and bold red lips. The face appears fragmented, with shards of glass or paper overlapping it, creating a sense of chaos. The eyes are intricately detailed, with the iris and pupil clearly visible, adding depth to the face. The striking red lips contrast with the predominantly dark and monotone background, which consists of grays, whites, and splashes of red. This cinematic piece exudes a dark fantasy vibe, drawing viewers in with its chaotic beauty, dark fantasy, illustration, cinematic, vibrant",
      "use_init_image": false,
      "use_lora": false,
      "seed": 1
    }
  }'

Response Format#

{
  "task": {
    "_id": "task_12345",
    "createdAt": "2024-01-29T21:52:07.171578+00:00",
    "updatedAt": "2024-01-29T21:52:07.171608+00:00",
    "user": "user_id",
    "requester": "requester_id",
    "tool": "txt2img",
    "parent_tool": null,
    "output_type": "image",
    "args": {
      "prompt": "A captivating and vibrant illustration...",
      "use_init_image": false,
      "use_lora": false,
      "seed": 1
    },
    "mock": false,
    "cost": 1,
    "handler_id": "fc-xxxxx",
    "status": "pending",
    "error": null,
    "result": null,
    "performance": null
  }
}

Common Issues#

Generation Quality#

Issue: Blurry or low-quality outputs
Solution: Use recommended resolutions, increase steps

Style Transfer#

Issue: Weak style application
Solution: Increase ipadapter_strength

Shape Guidance#

Issue: Poor shape adherence
Solution: Try different preprocessor types, adjust controlnet_strength

Use Cases#

Creative Projects#

Digital artwork creation
Concept art development
Character design
Illustration projects

Professional Applications#

Marketing materials
Product visualization
Social media content
Design mockups

Personal Use#

Avatar creation
Wallpaper generation
Gift artwork
Creative exploration

Tips for Best Results#

Prompt Engineering#

Be specific about style and medium
Include quality descriptors
Use clear, descriptive language
Reference specific artists or styles

Parameter Optimization#

Start with default settings
Adjust steps for quality vs. speed
Experiment with different seeds
Use appropriate resolutions

Next Steps#

Once you have your generated image:

Review Quality - Check for artifacts or issues
Adjust Parameters - Fine-tune for better results
Generate Variations - Create multiple samples
Export and Use - Save in your preferred format

Ready to create images? Visit the Text to Image tool to get started!