Generate Your First Image in 3 Steps

Step 1: Write Your Prompt

Describe the image you want to create. Be specific about style, lighting, composition, and subject.

Good vs Bad Prompts

Good	"a photorealistic golden retriever playing in autumn leaves, warm sunlight, bokeh background, 4K, detailed fur"
Bad	"dog"
Good	"cyberpunk city at night, neon signs, rain reflections on wet streets, cinematic lighting, ultra detailed"
Bad	"city"

Tip: Include style keywords (photorealistic, anime, oil painting), lighting (golden hour, dramatic), and quality tags (4K, detailed, high quality).

Step 2: Send Your Request

Use the API or the web playground to generate your image.

Option A: API (cURL)

curl -X POST https://api.arliai.com/v1/txt2img \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "prompt": "a photorealistic golden retriever playing in autumn leaves, warm sunlight, bokeh background",
    "negative_prompt": "blurry, low quality, distorted",
    "steps": 25,
    "sampler_name": "Euler a",
    "width": 1024,
    "height": 768,
    "cfg_scale": 7
  }' | python3 -c "
import sys, json, base64
data = json.load(sys.stdin)
img = base64.b64decode(data['images'][0])
with open('output.png', 'wb') as f:
    f.write(img)
print('Saved output.png')
"

Option B: API Gateway (No Auth)

curl -X POST https://api.kim8.s4s.host/api/v1/images/generations \
  -H "Content-Type: application/json" \
  -d '{
    "model": "stable-diffusion-xl-base-1.0",
    "prompt": "a photorealistic golden retriever playing in autumn leaves",
    "n": 1,
    "size": "1024x1024"
  }'

Option C: Web Playground

Visit generate.kim8.s4s.host, select a model, type your prompt, and click Generate.

Step 3: View Your Image

The API returns base64-encoded image data:

{
  "images": ["iVBORw0KGgoAAAANSUhEUgAA... (base64 PNG)"],
  "info": {
    "prompt": "a photorealistic golden retriever...",
    "seed": 1234567890,
    "steps": 25,
    "width": 1024,
    "height": 768
  }
}

Decode the base64 to save as a file. The playground displays images directly in your browser.

Next Steps

Image to Image	Transform existing images with a text prompt	Playground Guide
Upscale	Increase image resolution by 2x-4x	Upscale Guide
Model Guide	Find the best model for your use case	Model Guide
Full API Reference	All endpoints, parameters, and examples	API Reference

Quick Tips

• Start with FLUX.2-klein-4B or SDXL for general use
• Use steps: 20-30 for good quality without being slow
• Use cfg_scale: 7 as a starting point, increase for more prompt adherence
• Always include a negative prompt: "blurry, low quality, distorted, deformed"
• Note the seed from images you like for reproducible results