OmniGen V2

Edit images, swap outfits, personalize content, or create multi-person scenes with up to 3 images.

Input

Original

Output

Generated

Instructions

"Make the dress blue"

Upload your image and transform it in seconds

12,000+ images created this month

📄 About OmniGen V2

OmniGen V2 is a cutting-edge, unified multimodal AI model designed to revolutionize image editing and generation workflows. Leveraging state-of-the-art diffusion technology, OmniGen V2 enables users to perform advanced image manipulations, virtual try-ons, and multi-person compositions with unprecedented flexibility and control. Capable of working with up to three input images, this model excels at both simple edits and complex composite generation, making it an ideal tool for creatives and professionals seeking powerful, intuitive image editing solutions. At its core, OmniGen V2 combines separate text and image guidance, allowing users to specify detailed instructions for image edits or creations. For example, you can prompt the model to "Add bird from image 1 to desk in image 2" or "Change the color of the dress to blue." The dual-guidance system ensures that results are closely aligned with both the visual input and the textual prompt, delivering highly personalized and accurate outcomes. The model supports various image aspect ratios and resolutions, from square to portrait and landscape formats, with customizable image sizes ranging from 1024 to 4096 pixels. OmniGen V2 offers advanced control over the generation process through adjustable inference steps (20-50), guidance scales for both text and images, and configurable classifier-free guidance (CFG) range values. These features empower users to fine-tune quality, fidelity, and style according to their specific needs. The inclusion of two powerful diffusion schedulers—Euler and DPMSolver—further enhances creative control, enabling users to experiment with different generation dynamics and visual effects. Safety and quality are integral to OmniGen V2. The model incorporates an optional content safety checker and advanced negative prompt settings to minimize unwanted elements, such as deformities or artifacts. Users can also select their preferred output format (JPEG or PNG) and generate up to four images per request, supporting efficient batch creation and comparison. With random seed control for reproducibility and a robust synchronization mode, OmniGen V2 ensures consistent, reliable performance for a wide range of applications. Ideal use cases for OmniGen V2 include professional photo editing, content creation, e-commerce virtual try-on solutions, creative artwork generation, marketing collateral design, and collaborative multi-person image synthesis. Whether you are a designer, marketer, content creator, or developer, OmniGen V2 streamlines complex image tasks and opens new avenues for visual storytelling and personalization. Its pay-as-you-go credit system ensures scalable, flexible access to advanced AI-powered image editing capabilities without upfront commitments. Experience the next generation of multimodal image editing and unlock new creative possibilities with OmniGen V2.

✨ Key Features

Unified multimodal model supports image editing, generation, personalization, and virtual try-on capabilities.

Accepts up to 3 input images for advanced compositing and multi-person generation.

Separate text and image guidance allows precise control over edits and creative outputs.

Flexible image size options from 1024 to 4096 pixels, with multiple aspect ratios including square, portrait, and landscape.

Adjustable inference steps, guidance scales, and CFG range for fine-tuned quality and style.

Choice of Euler and DPMSolver diffusion schedulers for diverse generation dynamics.

Built-in safety checker and negative prompt settings to help avoid unwanted artifacts or inappropriate content.

💡 Use Cases

⚡Personalizing fashion or product images with virtual try-on and color changes.

⚡Editing and enhancing portraits by combining elements from multiple images.

⚡Creating marketing visuals and social media graphics with custom edits and enhancements.

⚡Generating creative artwork or illustrations from detailed prompts and image references.

⚡Building e-commerce catalogs with automated product swaps and background modifications.

⚡Composing multi-person group photos or scenes for creative projects.

⚡Developing content for advertising campaigns with unique visual concepts.

🎯 Best For

🎯 Professional designers, marketers, e-commerce teams, content creators, and developers seeking advanced AI-powered image editing and generation.

👍 Pros

✓Highly flexible with support for multiple input images and detailed prompts.

✓Advanced control over output quality and style through adjustable parameters.

✓Wide range of supported image sizes and aspect ratios for various applications.

✓Dual guidance system ensures outputs match both text and visual intent.

✓Efficient batch generation with up to four images per request.

✓Optional safety features reduce the risk of inappropriate or unwanted content.

⚠️ Considerations

△Requires careful prompt and parameter tuning for optimal results.

△Supports a maximum of three input images per generation.

△Generation times may vary based on complexity and image size.

△Some advanced features may require a learning curve for new users.

📚 How to Use OmniGen V2

Prepare your input images (up to three) and upload them to the platform.

Enter a detailed prompt describing the desired edit or image generation.

Select your preferred image size and aspect ratio from the available options.

Adjust advanced settings such as inference steps, guidance scales, and scheduler as needed.

Set a negative prompt if you want to avoid specific unwanted elements in the output.

Click generate and review the resulting images; download or further edit as desired.

Ready to try OmniGen V2?

Get 10 free credits — no credit card required

Start Free →

Frequently Asked Questions

OmniGen V2 can handle a wide range of image edits, including color changes, object additions or removals, virtual try-ons, and multi-person compositions. It supports both simple and complex transformations guided by text prompts and reference images.

You can upload and use up to three input images per generation. This allows for advanced compositing, multi-person edits, and the combination of elements from multiple sources.

Yes, you can adjust the text and image guidance scales to control how closely the output follows your prompt and input images. Fine-tuning these parameters helps achieve the desired level of accuracy.

Pricing varies by model and is based on a pay-as-you-go credit system. This approach provides flexibility and scalability without upfront costs, so you only pay for what you use.

You can choose between JPEG and PNG output formats for your generated images. This flexibility ensures compatibility with a wide range of applications and publishing needs.