Generate Your First Image

Step-by-step guide to generating your first image.

Prerequisites

• ComfyUI is installed
• A checkpoint model is downloaded
• Stable Diffusion model in ComfyUI/models/checkpoints/

Basic Steps

Step 1: Start ComfyUI

python main.py

Open browser: http://127.0.0.1:8188

Step 2: Load a Checkpoint

1. Right-click on canvas
2. Search "CheckpointLoader"
3. Click to add node
4. Select your model from dropdown

Step 3: Enter Positive Prompt

1. Right-click → "CLIP Text Encode"
2. Connect: Checkpoint → CLIP
3. Enter your prompt in text box

Example prompts:

# Portrait
a beautiful woman, detailed face, soft lighting, 8k

# Landscape
mountain landscape at sunset, golden hour, detailed

# Fantasy
magical forest, floating islands, fantasy art style

Step 4: Enter Negative Prompt

1. Add another "CLIP Text Encode" node
2. Connect to negative slot of KSampler
3. Enter things to avoid

Common negatives:

blurry, low quality, distorted, ugly, bad anatomy

Step 5: Set Sampler

1. Right-click → "KSampler"
2. Connect: Checkpoint → positive CLIP
3. Connect: Checkpoint → negative CLIP
4. Set parameters:

Parameter	Recommended	Description
Steps	20-30	Sampling steps
CFG	7-8	Prompt influence
Sampler	euler	Algorithm
Denoise	1.0	Full denoise

Step 6: Decode and Save

1. Add "VAEDecode" node
2. Connect KSampler → VAEDecode
3. Add "SaveImage" node
4. Connect VAEDecode → SaveImage

Step 7: Generate

Click "Queue Prompt" button (top right)

Complete Workflow

[CheckpointLoader]
      ↓
[CLIPTextEncode(positive)] → [KSampler] → [VAEDecode] → [SaveImage]
      ↑
[CLIPTextEncode(negative)]

Parameters Explained

Steps

Value	Effect
10-15	Fast, lower quality
20-30	Balanced
35-50	High quality, slower

CFG Scale

Value	Effect
3-5	More creative
6-8	Balanced
9-12	Strict following

Resolution

Size	Use Case
512x512	Quick preview
768x768	Standard
1024x1024	High quality

Tips

Better Results

• Use detailed prompts
• Add quality tags: "masterpiece, best quality"
• Use appropriate negative prompts
• Start with 512x512 for faster iteration

Common Issues

Issue	Solution
Black image	Check model file
Slow generation	Lower steps or use fp16
Out of memory	Reduce resolution

Next Steps

• Prompt Basics - Write better prompts
• Share Models - Share models
• Inpainting - Fix specific areas