> ## Documentation Index
> Fetch the complete documentation index at: https://docs.comput3.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Generate Medias Introduction
> Create images, videos, and multimedia content using AI on Comput3 Network
Harness the power of AI to generate high-quality images, videos, and multimedia content using ComfyUI on Comput3's distributed GPU infrastructure.
## Overview
Comput3 Network provides GPU instances pre-configured with ComfyUI for media generation:
Launch a GPU instance with ComfyUI pre-installed to create stunning images using Stable Diffusion workflows.
Use ComfyUI's video workflows on GPU instances to generate dynamic videos and animations.
Process audio files with WhisperX Gradio Interface for transcription, translation, and analysis.
Create and share custom ComfyUI workflows for specialized media generation tasks.
High-performance GPU instances optimized for ComfyUI and media generation workloads.
## Supported AI Models
### Image Generation Models
**Exceptional multilingual text rendering and editing capabilities**
* **Model**: 20B MMDiT model
* **Specialty**: Multilingual text rendering, advanced editing
* **Best for**: Text-heavy images, multilingual content, detailed editing
**High-quality unified multimodal models**
* **Text to Image**: 7B unified model with dual-path architecture
* **Image Edit**: Natural language image editing with text rendering support
* **Best for**: High-quality generation, versatile editing applications
**Physically accurate, high-fidelity image generation**
* **Model**: Cosmos-Predict2 2B T2I
* **Specialty**: Physically accurate, detail-rich generation
* **Best for**: Realistic images, scientific visualization, detailed artwork
**Multiple specialized models for different use cases**
* **I1 Dev**: Development and testing
* **I1 Rapide**: Fast image generation
* **I1 Complet**: Full-featured generation
* **E1.1 Image Edit**: Advanced editing (better quality than E1)
* **E1 Image Edit**: Standard image editing
**Latest Stable Diffusion with advanced features**
* **Simple**: Standard text-to-image generation
* **Grand Canny ControlNet**: Edge detection guided generation
* **Grande Profondeur**: Depth-aware image generation
* **Grand Flou**: Blur-based reference image generation
**High-quality SDXL models with various capabilities**
* **Simple**: High-quality standard generation
* **Refiner Prompt**: Enhanced results with refiners
* **Révisions de Texte**: Reference image concept transfer
* **Révision Zéro Positive**: Text prompts with reference images
* **Turbo**: Single-step image generation
**Modified Flux architecture for enhanced generation**
* **Model**: Chroma (modified from Flux)
* **Architecture**: Enhanced Flux-based architecture
* **Best for**: High-quality generation, architectural improvements
**Zero-shot monocular depth estimation**
* **Model**: Lotus Depth in ComfyUI
* **Specialty**: Efficient depth estimation with high detail retention
* **Best for**: Depth-aware applications, 3D processing
### Audio Processing Models
**Advanced audio transcription and analysis**
* **Model**: WhisperX with Gradio web interface
* **Features**: Speech-to-text, translation, speaker diarization, word-level timestamps
* **Languages**: 99+ languages supported
* **Best for**: Audio transcription, podcast processing, meeting notes, content analysis
### Video Generation Models
**Text-to-video generation**
* **Duration**: 4-16 seconds
* **Resolution**: 1280x768, 768x1280, 1024x1024
* **Frame Rate**: 24 FPS
* **Cost**: \$0.50 per 4-second clip
**Best for**: Social media content, marketing videos, creative projects
**Image-to-video animation**
* **Duration**: 2-8 seconds
* **Resolution**: 1024x576
* **Frame Rate**: 6-24 FPS
* **Cost**: \$0.25 per 4-second clip
**Best for**: Animating still images, product demonstrations, cinemagraphs
**Creative video generation**
* **Duration**: 3-12 seconds
* **Resolution**: Multiple aspect ratios
* **Features**: Camera controls, style consistency
* **Cost**: \$0.40 per 4-second clip
**Best for**: Creative storytelling, experimental videos, artistic content
## Getting Started
Launch a GPU instance with ComfyUI pre-installed from the Comput3 dashboard.
Use the ComfyUI template for instant setup with popular workflows pre-loaded.
Launch any GPU instance and install ComfyUI manually for full customization.
Connect to your GPU instance and open the ComfyUI web interface.
```bash theme={null}
# SSH into your instance
ssh -i your-key.pem ubuntu@
# ComfyUI should be running on port 8188
# Access via: http://:8188
```
Choose from pre-installed workflows or create your own:
* Text-to-image workflows
* Image-to-image transformations
* Video generation pipelines
* Custom node combinations
Run your workflows to generate images, videos, and other media content.
ComfyUI's node-based interface allows for complex, reproducible workflows.
## ComfyUI Interface
### Pre-configured Templates
Launch GPU instances with ComfyUI templates optimized for different use cases:
**Pre-loaded with popular image workflows**
* Stable Diffusion XL models
* ControlNet nodes pre-configured
* Popular LoRA models included
* Upscaling and enhancement workflows
**Specialized for video generation workflows**
* AnimateDiff models and workflows
* Stable Video Diffusion setup
* Frame interpolation nodes
* Temporal consistency tools
### ComfyUI Advantages
**Visual workflow creation and management**
* Drag-and-drop interface for building generation pipelines
* Reusable workflow components and templates
* Visual debugging and parameter adjustment
* Complex multi-step generation processes
**Support for any Stable Diffusion compatible model**
* Load custom checkpoints and LoRA models
* Switch between different model architectures
* Fine-tuned model integration
* Community model ecosystem access
**Precise control over generation parameters**
* ControlNet integration for guided generation
* Custom sampling methods and schedulers
* Batch processing and queue management
* Real-time parameter adjustment
### Getting ComfyUI Running
Use the ComfyUI template from the GPU instance launcher for instant setup.
For custom setups, install ComfyUI on any GPU instance:
```bash theme={null}
# Clone ComfyUI repository
git clone https://github.com/comfyanonymous/ComfyUI.git
cd ComfyUI
# Install dependencies
pip install -r requirements.txt
# Download base models
wget -O models/checkpoints/sd_xl_base_1.0.safetensors \
"https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/resolve/main/sd_xl_base_1.0.safetensors"
# Start ComfyUI server
python main.py --listen 0.0.0.0 --port 8188
```
## Prompt Engineering
### Effective Prompt Structure
**Format**: `[Subject] [Action] [Environment] [Style] [Technical]`
**Example**:
```
A majestic golden retriever running through a sunlit meadow,
impressionist painting style, soft lighting, high detail, 8k resolution
```
**Weight Control**:
```
(golden retriever:1.3) running through (sunlit meadow:0.8),
impressionist style, (soft lighting:1.2)
```
**Negative Prompts**:
```
Negative: blurry, low quality, distorted, ugly, bad anatomy
```
**Artistic Styles**:
* `in the style of Van Gogh`
* `digital art, concept art`
* `photorealistic, hyperrealistic`
* `anime, manga style`
* `oil painting, watercolor`
**Technical Modifiers**:
* `8k, ultra high resolution`
* `professional photography`
* `cinematic lighting`
* `depth of field, bokeh`
* `HDR, vibrant colors`
### Prompt Libraries
**Portrait Photography**:
* `professional headshot, studio lighting, shallow depth of field`
* `candid portrait, natural lighting, 85mm lens`
* `fashion photography, dramatic lighting, high contrast`
**Landscape Photography**:
* `golden hour landscape, warm lighting, wide angle`
* `dramatic mountain vista, storm clouds, HDR`
* `serene lake reflection, misty morning, soft colors`
**Digital Art**:
* `concept art, matte painting, detailed environment`
* `character design, game art, stylized rendering`
* `sci-fi illustration, futuristic, neon colors`
**Traditional Art**:
* `oil painting, classical style, renaissance lighting`
* `watercolor illustration, soft edges, flowing colors`
* `ink drawing, line art, minimalist style`
**Product Photography**:
* `product shot, white background, studio lighting`
* `lifestyle product photo, natural setting, soft shadows`
* `e-commerce photo, clean background, even lighting`
**Marketing Content**:
* `social media post, trendy style, vibrant colors`
* `banner design, modern layout, corporate style`
* `advertisement photo, professional, eye-catching`
## Quality and Resolution Options
### Image Resolutions
| Resolution | Aspect Ratio | Use Case | Cost Multiplier |
| ---------- | ------------ | ------------------------- | --------------- |
| 512x512 | 1:1 | Social media, avatars | 1x |
| 768x768 | 1:1 | High-quality social media | 1.5x |
| 1024x1024 | 1:1 | Print, detailed artwork | 2x |
| 1024x768 | 4:3 | Traditional photography | 2x |
| 1344x768 | 16:9 | Widescreen, banners | 2.5x |
| 1536x1024 | 3:2 | Professional photography | 3x |
### Quality Settings
**Settings**:
* Steps: 20-25
* Guidance Scale: 6-8
* Generation Time: 1-2 seconds
**Best for**: Rapid prototyping, concept exploration, batch generation
**Settings**:
* Steps: 30-35
* Guidance Scale: 7-9
* Generation Time: 2-4 seconds
**Best for**: General purpose, most use cases, good quality/speed balance
**Settings**:
* Steps: 40-50
* Guidance Scale: 8-12
* Generation Time: 4-8 seconds
**Best for**: Final artwork, print materials, professional use
## Pricing and Usage
### Cost Structure
Pricing is based on model complexity, resolution, and generation time. All prices include GPU compute and storage.
**Image Generation**:
* **SD 2.1**: \$0.01 per 512x512 image
* **SDXL**: \$0.02 per 1024x1024 image
* **DALL-E Style**: \$0.04 per 1024x1024 image
**Video Generation**:
* **Basic**: \$0.25 per 4-second 720p clip
* **HD**: \$0.50 per 4-second 1080p clip
* **Premium**: \$1.00 per 4-second 4K clip
### Usage Optimization
Generate multiple variations in a single request to reduce overhead costs.
Start with lower resolutions for iteration, then upscale final versions.
Choose the most cost-effective model that meets your quality requirements.
Refine prompts to reduce the number of generation attempts needed.
## Use Cases and Applications
### Content Creation
* **Social Media**: Instagram posts, Twitter headers, profile pictures
* **Marketing**: Ad banners, product mockups, promotional materials
* **Web Design**: Hero images, backgrounds, UI elements
* **Print Media**: Posters, flyers, book covers, magazine layouts
### Entertainment
* **Game Development**: Concept art, textures, character designs
* **Film/Video**: Storyboards, concept frames, visual effects plates
* **Animation**: Background art, character references, style guides
* **Virtual Reality**: Environment textures, 360° backgrounds
### Business Applications
* **E-commerce**: Product variations, lifestyle shots, catalog images
* **Real Estate**: Staged interiors, architectural visualizations
* **Fashion**: Model shots, product combinations, style guides
* **Education**: Illustrations, diagrams, educational content
### Creative Projects
* **Art Exploration**: Style experiments, creative inspiration
* **Personal Projects**: Custom artwork, gifts, decorations
* **Portfolio Work**: Concept pieces, artistic studies
* **NFT Creation**: Unique digital art, collectible series
## Next Steps
Learn detailed techniques for creating stunning AI-generated images.
Explore video creation with AI models and animation techniques.
Complete API documentation for programmatic media generation.