Supermachine AI Image Generation for E-Learning

Generate professional AI images for course assets using Supermachine.art - an AI image generator with 70+ models based on Stable Diffusion, SDXL, and FLUX.

Quick Reference

Platform: https://supermachine.art Documentation: https://docs.supermachine.art Image Generation Time: Under 15 seconds Commercial Use: Yes, all generated images

Core Workflow

1. Define the Image Need

Before generating, clarify:

•Purpose: Thumbnail, illustration, diagram background, hero image
•Style: Photorealistic, illustration, flat design, isometric, infographic
•Subject: People, objects, concepts, abstract
•Mood: Professional, friendly, energetic, calm
•Aspect Ratio: 16:9 (video), 1:1 (social), 4:3 (presentation)

2. Choose the Right Model

For Photorealistic Images (People, Offices, Workplaces):

•Flux - Best for realistic images WITH text capability
•Crystal Clear XL - Very realistic images
•Epic Realism - High-quality realistic photos
•Super Portrait - Professional headshots and portraits
•Models tagged "Photorealistic"

For Illustrations & Graphics:

•Supermachine Anime/Comic/Manga - Stylized illustrations
•Logos & Icons - Icon and logo generation
•Vector art models - Clean graphics
•Digital art models - Stylized visuals

For Text in Images:

•Flux - Only model that reliably renders text

3. Write Effective Prompts

Prompt Structure:

code

[Subject], [Style], [Details], [Lighting], [Camera/Composition], [Quality modifiers]

Example Prompts for Course Assets:

Professional Headshot for Instructor:

code

professional business portrait of a confident woman in her 30s,
wearing professional blazer, warm smile, office background with
bokeh, natural lighting, high quality, photorealistic

Course Thumbnail:

code

young professional looking at laptop with excited expression,
modern office, bright natural lighting, 16:9 aspect ratio,
photorealistic, professional stock photo style

Concept Illustration - Job Interview:

code

job interview scene, candidate and hiring manager shaking hands,
bright modern office, confident body language, professional
attire, natural lighting, photorealistic

Flat Design Illustration:

code

flat design illustration of resume document with magnifying glass,
minimalist style, professional blue color scheme, clean vectors,
white background

Infographic Background:

code

abstract gradient background, soft blue and white, professional,
minimal, clean, suitable for text overlay

4. Configure Settings

Essential Settings:

Setting	Recommended Value	Purpose
Model	Based on image type	See model selection above
Prompt Guidance	7-9	Balance creativity vs. accuracy
Steps	Default (model-specific)	Quality iterations
Resolution	Up to 1280×1280	Higher = more detail
Aspect Ratio	Match delivery format	16:9, 1:1, 4:3, etc.

Prompt Guidance Scale:

•5-6: More creative, may deviate from prompt
•7-9: Balanced (recommended for most uses)
•10-12: Strict adherence to prompt

5. Use Negative Prompts

Standard Negative Prompt for Professional Images:

code

poorly drawn, bad anatomy, wrong proportions, extra limbs,
cloned face, disfigured, gross proportions, malformed limbs,
missing arms, missing legs, extra arms, extra legs, fused fingers,
too many fingers, long neck, blurry, low quality, watermark,
text, signature

For Realistic Faces:

code

poorly drawn face, bad face, fused face, ugly face, worst face,
asymmetrical, unrealistic skin texture, double face

For Hands:

code

extra digits, extra hands, fused fingers, malformed limbs,
mutated hands, poorly drawn hands, extra fingers, missing hands,
bad hands, three hands, too many fingers, deformed hands

Advanced Features

ControlNet (Stable Diffusion models only)

Control composition and poses with reference images:

Type	Use Case
Canny	Match edge composition from reference
Open Pose	Control body position/orientation
Super Clone	Replicate specific person from photo
Composition	Arrange objects like reference

LoRA Models

Fine-tune specific attributes:

•Hair LoRA: -8 (straight) to 8 (maximum curl)
•Logo Design LoRA: Specific brand aesthetics
•Custom LoRAs: Per-model availability

Seed Numbers

•Use same seed to replicate exact image
•Find seeds in Gallery for images you like
•Useful for creating consistent character across images

E-Learning Use Cases

Course Thumbnails

code

Prompt: excited young professional celebrating at computer,
confetti, bright modern office, success moment, 16:9,
photorealistic, high energy
Model: Flux or Epic Realism
Aspect: 16:9

Module Header Images

code

Prompt: professional workspace with laptop and coffee,
resume on screen, morning light, shallow depth of field,
photorealistic
Model: Crystal Clear XL
Aspect: 16:9 or 3:1 for banners

Concept Illustrations

code

Prompt: flat design illustration of networking concept,
people connected by lines, diverse professionals,
modern style, blue and orange color scheme
Model: Vector or illustration model

Profile/Avatar Images

code

Prompt: professional headshot, diverse business person,
friendly expression, neutral background, studio lighting
Model: Super Portrait
Aspect: 1:1

Infographic Backgrounds

code

Prompt: abstract minimal gradient, soft professional colors,
clean, suitable for text overlay, corporate style
Model: Any SDXL model

Post-Processing Tools

After generating, use built-in tools:

Tool	Purpose
Remove BG	Transparent background for overlays
Upscale 4X	Increase resolution with detail enhancement
Face Swap	Replace faces (use ethically)
Image-to-Prompt	Analyze existing image for prompt ideas

Workflow for Course Development

Batch Generation Strategy

•Create Prompt Templates for each image type needed
•Generate Variations - Create 3-4 options per concept
•Note Successful Seeds - Reuse for consistency
•Upscale Final Selections - Use 4X upscale for print/high-res

Maintaining Visual Consistency

•Use same model for similar image types
•Keep consistent prompt style elements
•Note and reuse seeds for character consistency
•Maintain color scheme keywords across prompts

Credit Management

Plan	Credits	Best For
Starter ($5)	100	Testing
Apprentice ($19/mo)	1,000	Light use
Master ($49/mo)	4,000	Heavy development

Prompt Library for Job Search Course

Resume & Documents

code

professional resume document on desk, pen beside it,
soft natural lighting, shallow depth of field, clean
modern aesthetic, photorealistic

Interview Scenes

code

confident job candidate in interview, positive body
language, professional attire, modern conference room,
natural window lighting, photorealistic

Networking

code

professionals networking at event, diverse group,
friendly conversation, modern venue, warm lighting,
photorealistic

Job Search

code

person using laptop for job search, focused expression,
coffee shop or home office, natural lighting,
contemporary setting, photorealistic

Success/Celebration

code

young professional receiving job offer, excited expression,
holding offer letter, bright natural light, joyful moment,
photorealistic

LinkedIn Profile

code

professional headshot suitable for LinkedIn, confident
smile, business casual, neutral background, soft studio
lighting, photorealistic, approachable

Quality Checklist

Before using generated images:

• Image matches intended purpose and mood
• No visible AI artifacts (extra fingers, distorted faces)
• Appropriate for target audience (diverse, relatable)
• Resolution sufficient for delivery format
• Consistent with course visual style
• Background suitable for text overlay (if needed)
• Proper licensing for commercial e-learning use

Tips for Best Results

•Be Specific: More detail = better results
•Use Quality Modifiers: "high quality, professional, 8k, detailed"
•Specify Lighting: "natural light, studio lighting, soft lighting"
•Include Mood: "confident, friendly, professional, energetic"
•Iterate: Generate multiple versions, refine prompts
•Check Gallery: Find inspiration and working prompts
•Use Prompt Book: Pre-made prompts for quick starts
•Match Model to Task: Photorealistic vs illustration needs different models

Resources

•Supermachine Gallery: See what prompts create which results
•Prompt Book: Pre-made prompt templates within platform
•Style Assistance: Built-in tools for emotions, lighting, camera settings
•Model Tags: Use "Photorealistic" tag for realistic needs

Troubleshooting

Poor Image Quality

•Increase steps (if available)
•Use higher resolution
•Add quality modifiers to prompt
•Try different model

Faces Look Wrong

•Add negative prompt for face issues
•Use portrait-specific model
•Reduce prompt guidance slightly

Hands Look Distorted

•Add hand-specific negative prompt
•Crop or position hands out of frame
•Use post-editing if needed

Text Not Rendering

•Use Flux model (only one with text capability)
•Keep text short and simple
•Specify text clearly in quotes

Inconsistent Style

•Use same model throughout project
•Create and reuse prompt templates
•Save and reuse successful seeds