Most advanced model offering exceptional quality up to 4 megapixels, featuring Ultra and Raw modes for superior detail and natural aesthetics.

High-performance model delivering professional-grade outputs with excellent image quality and strong prompt adherence.

Flagship model offering state-of-the-art performance with top-tier prompt following and visual quality.

State-of-the-art inpainting and outpainting model for seamless image editing and extension.

Professional depth-aware model enabling superior 3D understanding and spatial composition.

Advanced edge-detection model for precise structural control.

Open-weight model balancing quality and performance, designed for non-commercial applications.

An adaptation of the FLUX Dev model that incorporates Low-Rank Adaptation (LoRA) technology. This allows for efficient customization of the model for specific styles or domains while maintaining high-quality outputs.

Development version of the inpainting model for testing and prototyping.

12-billion parameter model for depth-aware image generation.

Development version using ControlNet for Canny edge detection.

An adapter for the FLUX Dev model that allows for mixing and recreating input images with text prompts. It's particularly useful for image variation and restyling tasks.

High-speed model optimized for rapid generation, capable of producing images in 1-4 steps.

Similar to Redux Dev, but optimized for speed. It allows for quick image variations and restyling using the faster FLUX Schnell base model.

A version of the FLUX Schnell model that incorporates LoRA technology. This allows for rapid fine-tuning and adaptation of the fast Schnell model to specific styles or use cases while maintaining its speed advantage.

A specialized model designed to generate cinematic-style images with film-like qualities. It's ideal for creating movie-inspired visuals or scenes with a cinematic aesthetic.

An anime and illustration-style image generation model inspired by Studio Ghibli's distinctive art style. It excels at creating whimsical, fantastical scenes with a characteristic Ghibli-esque charm.

A model designed for high-energy, stylized image generation. It's capable of producing vibrant, dynamic visuals with exaggerated features or effects.

A model focused on shape and form detection. It's particularly useful for architectural and design applications, helping to identify and generate specific structural forms.

Fine-tuned model for creating surreal, otherworldly scenes with glowing, neon-like elements.

Recraft V2 and V3 are AI models designed for versatile image generation. These models offer a range of capabilities, from high-quality professional outputs to anatomically accurate designs, catering to diverse user needs in graphic design and creative projects.

Recraft V3 SVG and Recraft V2 SVG are AI models designed for versatile vector image generation. These models offer a range of capabilities, from high-quality professional SVG outputs to detailed vector graphics, catering to diverse user needs in logo design, icon creation, and scalable illustrations.

SDXL BASE 1.0 is a state-of-the-art text-to-image generation model developed by Stability AI. It features 3.5 billion parameters, dual text encoders for better prompt understanding, and generates images at a native resolution of 1024x1024 pixels. Known for its exceptional visual quality, it excels in photorealism, artistic styles, and handling complex compositions. SDXL also simplifies prompt requirements and supports multi-aspect ratios, making it versatile for various creative applications.

SDXL Pony is an advanced text-to-image model built on Stable Diffusion XL, adept at creating high-quality images across diverse subjects and styles, including humans, animals, and artistic genres. Trained on 2.6M images, it specializes in both SFW and NSFW outputs with strong natural language understanding, vibrant colors, and detailed compositions, making it ideal for fantasy and anime creators.

Create Custom Emoji Masterpieces: Inspired by Apple's Iconic Style

An advanced text-to-image model optimized for fast, high-quality results in 4 inference steps. Using Adversarial Diffusion Distillation (ADD), it delivers detailed, photorealistic images and supports diverse artistic styles. Ideal for applications like digital art, media production, and e-commerce.

An advanced text-to-image model with 8.1B parameters, generating up to 1MP high-quality images. Ideal for digital art, advertising, and VFX, it employs a Multimodal Diffusion Transformer for diverse outputs and exceptional photorealism, useful in media, gaming, and education.

A text-to-image model optimized for creating high-quality images quickly. Balances speed and detail, supporting diverse styles and uses like design, content creation, and prototyping.

This model introduced the Multimodal Diffusion Transformer (MMDiT) architecture, replacing the UNet with a Rectified Flow Transformer. This change significantly enhanced image quality and prompt adherence, offering various model sizes to balance performance and resource efficiency.

This version introduces latent diffusion for text-to-image synthesis, being capable of generating images at 512x512 pixels. It quickly gained popularity for its ability to create detailed and artistic visuals. It is now considered deprecated.

Image Restoration Algorithm for Blured/Bad images.

Face Restoration Algorithm for Blured/Bad images.

Swap faces.

Create a prompt based on an image.

Remove the background of an image.

Image UpScaling Algorithm.