Image Generation

Create beautiful, unique images using AI. The platform's Image Generation feature integrates the best image generation models of 2026 — Google Imagen 4, GPT Image 1.5, FLUX.2, and BytePlus Seedream — transforming your text descriptions into stunning visuals.

Getting Started

Access Image Generation

1. Sign in to the platform
2. Navigate to Agents in the sidebar
3. Select Image Generation

Create Your First Image

1. Choose an AI model (see Available Models below)
2. Type a description of the image you want to create
3. Optionally adjust settings (aspect ratio, style, quality tier, etc.)
4. Click Generate
5. Wait a few seconds for your image to appear
6. Download or save to your collection

Available Models

Google Imagen 4

Google's premier image generation family, available via Vertex AI and Gemini API:

All images include Google's SynthID watermark to identify AI-generated content. Imagen 4 offers the best price-to-quality ratio for most use cases.

GPT Image 1.5 (OpenAI)

OpenAI's current flagship image model, replacing DALL-E 3:

GPT Image 1.5 achieves the highest LM Arena Elo score of 1,264 among all tested image models as of early 2026, with 4× faster generation than DALL-E 3 and significantly better photorealism.

Note: DALL-E 3 remains available but is effectively legacy technology, superseded by GPT Image 1.5.

FLUX.2 (Black Forest Labs)

A powerful open-source-friendly alternative with pay-per-use pricing:

• FLUX.1.1 Pro — highest technical quality in the FLUX family, 4.5-second generation, ideal for commercial realism
• Strong on detail, anatomy, and complex scene composition
• Available via API without subscription lock-in

BytePlus Seedream

ByteDance's creative image generation model integrated into the desktop app:

• Optimized for artistic and stylized outputs
• Available via the desktop application's media generation tools

Writing Effective Prompts

Prompt Basics

The better your description, the better your results. Include:

• Subject — What is the main focus?
• Style — Photorealistic? Illustration? Painting?
• Mood — Bright? Dark? Peaceful? Energetic?
• Details — Colors, lighting, composition

Examples

Simple Prompt:

"A cat sitting on a windowsill"

Detailed Prompt:

"A fluffy orange tabby cat sitting on a sunny windowsill, looking out at a garden, soft natural lighting, cozy atmosphere, photorealistic"

Artistic Prompt:

"A futuristic cityscape at sunset, cyberpunk style, neon lights reflecting on wet streets, dramatic clouds, digital art"

Tips for Better Results

Settings & Options

Aspect Ratio

Choose the shape of your image:

Style Presets

Quick-start your creative direction:

• Photorealistic — Looks like a real photograph
• Digital Art — Clean, polished digital illustration
• Painting — Traditional art styles (oil, watercolor, etc.)
• Anime — Japanese animation style
• 3D Render — Three-dimensional CGI look
• Sketch — Hand-drawn, pencil or ink style

Advanced Settings

Negative Prompts Specify what you don't want in the image:

• "no text," "no watermarks"
• "not blurry," "no distortion"
• "no people," "without background"

Seed Number For reproducible results:

• Use the same seed to get similar images
• Change the seed for variations
• Share seeds to recreate specific styles

Number of Images Generate multiple variations at once:

• Compare different interpretations
• Choose the best result
• More efficient than generating one at a time

Quality Tier Select Low, Standard, or High depending on your needs:

• Low for quick concept exploration
• High for final production assets

Working with Results

Viewing Images

• Click any image to view full-size
• Use arrow keys to navigate between images
• Zoom in for detail inspection

Saving Images

1. Click the image to select it
2. Click Download for your device
3. Or click Save to Collection to keep in the platform

Image Collections

Organize your generated images:

• Create named collections
• Add favorites for quick access
• Search through past generations
• Share collections with team members

Regenerating

Not quite right? Try:

• Regenerate — Create new images with same prompt
• Vary — Make small variations of a specific image
• Edit Prompt — Refine your description and try again

Use Cases

Marketing & Social Media

• Social media graphics
• Blog post headers
• Ad creatives
• Email newsletter images

Design & Mockups

• Concept art and mood boards
• Product visualization
• Presentation graphics
• Website imagery

Creative Projects

• Illustrations for stories
• Character designs
• Landscape art
• Abstract backgrounds

Best Practices

Do's

✅ Be descriptive and specific ✅ Include style and mood keywords ✅ Generate multiple variations ✅ Use negative prompts to refine ✅ Save seeds for successful images ✅ Iterate and refine your prompts ✅ Use the appropriate quality tier for your needs

Don'ts

❌ Use single-word prompts ❌ Expect exact reproductions of copyrighted works ❌ Include text in images (AI still struggles with precise text rendering) ❌ Generate inappropriate or harmful content ❌ Claim AI images as photographs without disclosure ❌ Register copyright on purely AI-generated images (regulations vary by jurisdiction)

Usage & Limits

Subscription Limits

Check Your Usage

1. Go to Settings → Usage
2. View remaining generations
3. See usage history

Rate Limits

To ensure fair usage:

• Maximum 5 concurrent generations
• Brief cooldown between large batches
• Higher limits for paid plans

Troubleshooting

Generation Failed

• Try simplifying your prompt
• Check your internet connection
• Verify you have remaining generations
• Try again in a few moments

Unexpected Results

• Add more detail to your prompt
• Use negative prompts to exclude unwanted elements
• Try different style keywords
• Generate more variations
• Switch to a different model (e.g., try FLUX.2 for complex realism)

Slow Generation

• Complex prompts take longer
• High-resolution images need more time
• Imagen 4 Fast is the quickest option for speed
• Peak hours may have delays — check the status indicator

Content Blocked

Some prompts may be blocked for safety:

• Avoid violence, explicit content, or harmful imagery
• Real people/celebrities are restricted
• Copyrighted characters may be limited
• Rephrase and try again

Keyboard Shortcuts

References

• The 9 Best AI Image Generation Models in 2026 — gradually.ai
• AI Image Generation API Comparison 2026 — LaoZhang AI Blog
• Google Imagen 4 — Gemini API for Developers
• AI Image Generators: Complete Guide 2026 — AfricanAI
• AI Image Pricing 2026: Google Gemini vs. OpenAI GPT — IntuitionLabs
• Best AI Image Generators 2026 — Tom's Guide