Skip to main content
Generate engaging videos using ComfyUI’s video workflows on Comput3 Network’s GPU infrastructure. Launch a GPU instance with ComfyUI and access powerful video generation capabilities.

Quick Start

1

Launch GPU Instance

Launch a GPU instance with ComfyUI pre-configured for video generation.

Video Template

Use the ComfyUI Video template with video generation workflows pre-installed.

Recommended GPU

A100 or H100 recommended for video generation workflows due to high VRAM requirements.
2

Access ComfyUI Video

Connect to your instance and open ComfyUI with video generation capabilities:
# SSH into your instance
ssh -i your-key.pem ubuntu@<instance-ip>

# ComfyUI Video interface
http://<instance-ip>:8188
Video generation workflows require significant VRAM. Ensure your GPU instance has at least 16GB VRAM.
3

Load Video Workflow

Choose from available video generation workflows in ComfyUI:
AnimateDiff workflows for text-to-video generation
  • Stable Video Diffusion models
  • AnimateDiff motion modules
  • ControlNet for precise control
  • Custom motion prompts
4

Configure and Generate

Configure your workflow parameters in ComfyUI:
Essential settings for video generation
  • Prompt: Text description of your video
  • Frame Count: Number of frames to generate (24 = 1 second at 24fps)
  • Width/Height: Video resolution (1024x576, 512x512, etc.)
  • Seed: For reproducible results
  • Steps: Denoising steps (20-50 typical range)
Control movement and animation
  • Motion Strength: How much movement to add (0.5-1.5)
  • Motion Module: Choose animation style (smooth, dramatic, etc.)
  • Context Length: How many frames to consider for consistency
  • FPS: Target frame rate for output
Optimize for quality vs speed
  • Sampler: DPM++, Euler, or other sampling methods
  • CFG Scale: Guidance strength (6-12 typical)
  • Denoising Strength: For image-to-video workflows
  • Upscaling: Post-process to higher resolutions
5

Download and Use

Your generated video will be processed and available for download in MP4 format.
Video generation typically takes 2-5 minutes depending on duration and complexity.

Video Models and Capabilities

Text-to-Video Models

Industry-leading text-to-video generationSpecifications:
  • Duration: 4-16 seconds
  • Resolution: 1280x768, 768x1280, 1024x1024
  • Frame Rate: 24 FPS
  • Aspect Ratios: 16:9, 9:16, 1:1
Strengths:
  • Exceptional prompt following
  • Smooth, realistic motion
  • High visual quality and detail
  • Great for narrative content
Best Use Cases:
  • Marketing videos and ads
  • Social media content
  • Concept visualization
  • Storytelling and narrative clips
Pricing: $0.50 per 4-second segment
Creative and artistic video generationSpecifications:
  • Duration: 3-12 seconds
  • Resolution: 1024x768, 768x1024
  • Frame Rate: 12-24 FPS
  • Unique Features: Camera controls, style consistency
Strengths:
  • Creative and artistic interpretations
  • Unique visual styles and effects
  • Good motion dynamics
  • Experimental and avant-garde results
Best Use Cases:
  • Artistic projects and experiments
  • Creative storytelling
  • Music videos and artistic content
  • Experimental marketing
Pricing: $0.40 per 4-second segment
Open-source video generationSpecifications:
  • Duration: 2-8 seconds
  • Resolution: 1024x576, 576x1024
  • Frame Rate: 6-12 FPS
  • Models: SVD, SVD-XT
Strengths:
  • Cost-effective generation
  • Good for simple animations
  • Consistent with Stable Diffusion aesthetics
  • Customizable and extensible
Best Use Cases:
  • Simple animations and loops
  • Product demonstrations
  • Educational content
  • Budget-friendly projects
Pricing: $0.20 per 4-second segment

Image-to-Video Animation

Bring still images to life
{
  "model": "stable-video-diffusion",
  "input_image": "https://your-image.jpg",
  "motion_prompt": "gentle swaying in the wind",
  "duration": 6,
  "motion_strength": 0.7,
  "fps": 12
}
Animation Types:
  • Subtle Motion: Gentle movements, breathing, wind effects
  • Camera Movement: Zoom, pan, rotation effects
  • Object Animation: Moving elements within the scene
  • Atmospheric Effects: Flowing water, moving clouds, fire

Advanced Video Prompting

Effective Video Prompts

Describe the type and quality of movementCamera Movement:
  • “slow zoom in on the subject”
  • “smooth panning from left to right”
  • “orbital camera movement around object”
  • “handheld camera, slight shake”
  • “drone shot, rising upward”
Subject Movement:
  • “person walking slowly toward camera”
  • “leaves falling gently from tree”
  • “waves crashing against rocks”
  • “bird flying in slow motion”
  • “dancer spinning gracefully”
Control the speed and rhythm of motion
  • “slow motion, dreamy pace”
  • “fast-paced, energetic movement”
  • “gradual acceleration”
  • “rhythmic, pulsing motion”
  • “sudden, dramatic movement”
  • “gentle, flowing transitions”
Specify the aesthetic and mood
  • “cinematic quality, film grain”
  • “bright and vibrant colors”
  • “moody, dramatic lighting”
  • “soft, ethereal atmosphere”
  • “high contrast, bold shadows”
  • “vintage film aesthetic”
Describe the overall scene structure
  • “wide establishing shot of landscape”
  • “close-up focus on facial expressions”
  • “medium shot with shallow depth of field”
  • “bird’s eye view of the action”
  • “intimate, personal perspective”

Motion Control Parameters

Control the intensity of movement
{
  "motion_strength": 0.3,  // Subtle, gentle movement
  "motion_strength": 0.7,  // Moderate, noticeable motion
  "motion_strength": 1.0   // Strong, dramatic movement
}
Guidelines:
  • 0.1-0.3: Subtle effects (breathing, gentle sway)
  • 0.4-0.7: Moderate motion (walking, flowing)
  • 0.8-1.0: Dramatic action (running, dancing)

Video Enhancement and Post-Processing

Upscaling and Quality Enhancement

Enhance video resolution and quality
curl
curl -X POST "https://api.comput3.ai/v1/enhance/video-upscale" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "video_url": "https://your-video.mp4",
    "scale_factor": 2,
    "model": "real-esrgan-video",
    "enhance_faces": true
  }'
Options:
  • Scale factors: 2x, 4x
  • Face enhancement: Improve facial details
  • Noise reduction: Remove compression artifacts
  • Frame interpolation: Increase frame rate
Increase frame rate for smoother motion
{
  "video_url": "https://your-video.mp4",
  "target_fps": 60,
  "interpolation_model": "rife",
  "preserve_motion": true
}
Benefits:
  • Smoother motion appearance
  • Better slow-motion effects
  • Improved playback quality
  • Professional video standards
Reduce camera shake and improve stability
response = requests.post(api_url, json={
    "video_url": "https://shaky-video.mp4",
    "stabilization_strength": 0.8,
    "crop_to_stabilize": True,
    "smooth_motion": True
})

Style Transfer and Effects

Apply artistic styles to videos
{
  "video_url": "https://your-video.mp4",
  "style": "oil_painting",
  "strength": 0.7,
  "temporal_consistency": true,
  "preserve_faces": false
}
Available Styles:
  • Oil painting, watercolor, pencil sketch
  • Pop art, abstract, impressionist
  • Anime, cartoon, comic book
  • Film noir, vintage, retro

Video Formats and Export Options

Output Formats

Quality and Compression Settings

For professional use and archival
{
  "bitrate": "high",
  "compression": "minimal",
  "quality_preset": "professional"
}
  • Larger file sizes
  • Maximum detail preservation
  • Best for editing and professional use
Smallest file sizes for bandwidth-limited scenarios
{
  "bitrate": "low",
  "compression": "aggressive",
  "quality_preset": "mobile"
}
  • Smallest file sizes
  • May sacrifice some quality
  • Good for mobile and slow connections

Batch Video Generation

Generating Video Series

import requests
import time

def generate_video_series(prompts, model="runway-gen2"):
    results = []
    
    for i, prompt in enumerate(prompts):
        response = requests.post(
            "https://api.comput3.ai/v1/generate/video",
            headers={"Authorization": "Bearer YOUR_API_KEY"},
            json={
                "model": model,
                "prompt": prompt,
                "duration": 6,
                "width": 1280,
                "height": 768,
                "fps": 24
            }
        )
        
        result = response.json()
        results.append({
            "prompt": prompt,
            "video_id": result["id"],
            "status": result["status"]
        })
        
        # Wait between requests to avoid rate limiting
        time.sleep(2)
    
    return results

# Generate a series of related videos
story_prompts = [
    "A person walking through a forest path at dawn",
    "The same person discovering a hidden waterfall",
    "Close-up of wonder and amazement on their face",
    "Wide shot of the magnificent waterfall scene"
]

video_series = generate_video_series(story_prompts)

Video Concatenation and Editing

Combine multiple generated videos into one
{
  "video_urls": [
    "https://video1.mp4",
    "https://video2.mp4", 
    "https://video3.mp4"
  ],
  "transitions": [
    {"type": "fade", "duration": 0.5},
    {"type": "crossfade", "duration": 1.0}
  ],
  "output_format": "mp4"
}

Commercial Use and Pricing

Pricing Structure

Video generation pricing is based on duration, resolution, and model complexity. All prices include processing and storage.
Text-to-Video:
  • Runway Gen-2 Style: $0.50 per 4-second segment
  • Pika Labs Style: $0.40 per 4-second segment
  • Stable Video: $0.20 per 4-second segment
Image-to-Video:
  • Standard Animation: $0.25 per 4-second segment
  • Advanced Motion: $0.35 per 4-second segment
Video Enhancement:
  • Upscaling: $0.15 per 4-second segment
  • Style Transfer: $0.20 per 4-second segment
  • Stabilization: $0.10 per 4-second segment

Usage Rights

Included with all plansAllowed:
  • Personal projects and social media
  • Educational and research purposes
  • Portfolio and creative work
  • Non-commercial content creation
Available with Pro and Enterprise plansAllowed:
  • Marketing and advertising campaigns
  • Social media marketing
  • Product demonstrations
  • Client work and services
  • Broadcast and streaming content
Restrictions:
  • Attribution may be required
  • Some models have usage limitations
  • High-volume use may require special licensing
For television, film, and major mediaIncludes:
  • Television and film use
  • Streaming platform content
  • Major advertising campaigns
  • No attribution requirements
  • Extended usage rights
Contact sales for pricing and terms

Troubleshooting and Optimization

Common Issues

Symptoms: Jerky, unnatural, or inconsistent movementSolutions:
  • Adjust motion strength (try 0.5-0.8 range)
  • Use more specific motion descriptions
  • Try different models for different motion types
  • Increase duration for smoother motion
  • Use temporal consistency settings
Example Improvements: ❌ “Person moving” ✅ “Person walking slowly and smoothly toward camera”
Symptoms: Flickering, objects changing appearance between framesSolutions:
  • Enable temporal consistency options
  • Use more detailed and specific prompts
  • Avoid overly complex scenes
  • Try different sampling methods
  • Consider using image-to-video for more control
Symptoms: Videos taking longer than expected to generateSolutions:
  • Reduce video duration for faster processing
  • Lower resolution if quality allows
  • Use faster models for iteration
  • Generate during off-peak hours
  • Check system status for any delays
Symptoms: Generated video doesn’t match descriptionSolutions:
  • Be more specific about motion and camera work
  • Break complex prompts into simpler descriptions
  • Use reference images when available
  • Try different models that specialize in your content type
  • Iterate with variations of the prompt

Optimization Tips

Quality Optimization

  • Use detailed, specific prompts
  • Choose appropriate models for content type
  • Consider motion strength carefully
  • Use reference images when possible

Cost Optimization

  • Start with shorter durations for testing
  • Use appropriate resolution for final use
  • Batch similar requests together
  • Choose cost-effective models when quality allows

Speed Optimization

  • Use faster models for iteration
  • Generate during off-peak hours
  • Prepare detailed prompts in advance
  • Use batch processing for multiple videos

Workflow Optimization

  • Save successful prompts and settings
  • Plan video sequences in advance
  • Use storyboards for complex projects
  • Organize generated content with metadata

Creative Applications

Marketing and Advertising

  • Product Demos: Showcase products in action
  • Brand Storytelling: Create engaging brand narratives
  • Social Media Content: Generate eye-catching video posts
  • Explainer Videos: Visualize complex concepts

Entertainment and Media

  • Music Videos: Create artistic visuals for music
  • Short Films: Generate scenes and sequences
  • Game Trailers: Create promotional content
  • Documentary B-roll: Generate supporting footage

Education and Training

  • Educational Content: Visualize learning materials
  • Training Videos: Create instructional content
  • Scientific Visualization: Illustrate complex processes
  • Historical Recreation: Visualize historical events

Art and Creative Projects

  • Digital Art: Create moving art pieces
  • Experimental Film: Explore new visual styles
  • Interactive Installations: Generate dynamic content
  • NFT Collections: Create unique digital collectibles

Next Steps