​

MediaGen Guide

​

Quick Start

​

Basic Setup

1. Use /model to select your preferred model
2. Enter your private thread (auto-named with a numerical string)
3. Access media commands with /:

/text_to_video    # Generate videos from text
/text_to_image    # Create images from descriptions
/image_to_text    # Convert images to text descriptions

​

Available Models

​

Multimodal Support

• Video Processing: Nova Lite/Pro
• Video Generation: Reels/Ray2
• Image Generation: Canvas
• Vision Models: Claude Sonnet, Haiku 3, Llama 11b/90b

​

Working with Media

​

Video Generation (Reels)

/text_to_video    # Generate 5-9-second videos
/image_to_video   # Convert static images to videos

• Processing time: 5-6 minutes
• Character limit: 512 characters(reels) / 1,024 characters(ray2)

​

Image Generation (Canvas)

/text_to_image    # Create images from descriptions

• Processing time: 1-2 minutes
• Character limit: 1,024 characters

​

Video Understanding (Nova)

1. Upload your video
2. Provide content description
3. Specify desired output
4. Submit for processing

​

Vision Models (Claude/Llama)

1. Upload your image
2. Add content description
3. Specify analysis requirements
4. Wait for model response

​

Thread Management

​

Private Threads

• Messages are ephemeral (visible only to creator)
• Access threads via #model-chat
• Look for blinking dot in channel menu
• Right-click to view channel names
• Invite others using @ mentions

​

System Prompts

• Optional thread naming feature
• Defines model purpose/role
• Available for text/vision models
• Set via system_prompt parameter

​

Best Practices

​

Workflow Optimization

• Use multimodal models for context preparation
• Pre-prompt with image or text context
• Consider model-specific strengths
• Mind character limits for each type

​

Queue Management

• Requests processed in order
• Include cooldown periods
• Check queue status before submitting
• Use help commands for guidance

​

Support Options

​

Getting Help

• Direct message @StreamPod AI
• Visit #general-support channel
• Use / commands for help documentation
• Check model-specific guidelines

​

Common Solutions

• Respect character limits
• Allow sufficient processing time
• Use appropriate models for tasks
• Follow upload-then-prompt sequence

​

Advanced Tips

​

Model Selection

• Choose based on media type
• Consider processing time needs
• Review character limitations
• Check multimodal capabilities

​

Content Guidelines

• Provide clear descriptions
• Include relevant context
• Follow upload procedures
• Monitor processing status