Changelog

The GPT Image 2 API leverages OpenAI’s gpt-image-2 model to generate high-quality images from text prompts with customizable quality levels and sizes. It supports image-to-image editing via reference images, batch generation, and flexible aspect ratios.

The Virtual Try-On API enables realistic clothing transfer onto a model image. It supports upper garments, lower garments, and one-piece outfits, allowing you to visualize how different clothing items look on a person.

The Nano Banana Pro API harnesses the Gemini 3 Pro Image Preview model for premium, high-fidelity image generation from text prompts. It delivers superior image quality with support for 1K and 2K resolutions.

The Nano Banana 2 API leverages the Gemini 3.1 Flash Image Preview model to generate high-quality images from text prompts. It delivers fast, cost-efficient image generation with support for multiple resolutions up to 4K.

The DreamAPI Callback feature enables event-driven result delivery for all async API tasks. By adding callback parameters to your async request, DreamAPI will automatically send a POST request to your specified callback URL once the task is completed, eliminating the need for continuous polling.

The Video Enhance API upscales and improves video quality using advanced super-resolution algorithms. This API enhances video clarity, sharpness, and overall visual quality while preserving the original content.

The SeeDream API generates high-quality images from text prompts, supporting multiple model versions for different quality and speed requirements.

The Seedance 2.0 API is an advanced video generation model that supports multiple input types including text, images, videos, and audio for comprehensive video creation.

Video Translate 2.0 provides comprehensive end-to-end video localization. It automatically extracts audio, performs Automatic Speech Recognition (ASR) to generate source text, translates content via Machine Translation (MT), and utilizes voice cloning (TTS) with advanced lip-sync technology. The result is a new audio track with natural voice quality and perfectly synchronized lip movements. This version is optimized for single-speaker scenarios, making it ideal for high-quality multilingual adaptation of courses, lectures, and interviews.

The Lipsync 2.0 API is an asynchronous audio-video synchronization service that aligns specified audio with character lip movements in videos. This upgraded version significantly improves output video clarity and resolution compared to v1.0.

This API enables video-driven talking head generation, animating facial images to match the expressions, mouth movements, and head poses of a source driving video. The service reconstructs a drivable face model from input images and applies motion sequences extracted from the driving video to produce realistic animation.

The DreamAvatar 3.0 Fast API generates talking avatars from audio and image inputs. It accurately mirrors facial expressions and vocal tones of real individuals while supporting multiple languages.

Text To Video (Wan 2.1) automatically generates 4-second video clips based on text prompts and descriptions.

Image To Video (Wan 2.1) automatically generates 4-second video clips based on text prompts and uploaded source images.

Head Tail To Video (Wan 2.0) automatically generates 4-second videos based on a text prompt and provided head/tail frame images.

This API enables high-quality image generation using existing images and natural language descriptions. It leverages advanced AI models to produce professional visuals for design automation, e-commerce, advertising, and game development.

This API generates high-quality images directly from text prompts. It utilizes powerful AI models to interpret natural language and produce visuals for content generation, advertising, and creative design.

The PAG Merge API performs PAG-based video processing. It requires a PAG file URL and customization parameters to render videos with specified visual effects or dynamic content defined by the PAG template.

Retrieves the current balance of available credits for the account.

Scene Training allows users to upload images to train or fine-tune custom scene models. The system learns specific visual characteristics and patterns from the input to adapt to custom contextual environments.

Image To Video (Wan 2.0) generates 4-second videos based on text prompts and source images.

Text To Video (Wan 2.0) generates 4-second videos based on natural language descriptions.

The Image To Video (Sonic) API generates talking avatars from audio and images, mirroring realistic facial expressions and vocal tones with multilingual support.

The Demoire API removes screen-induced stripes and moiré patterns from images.

The Shadow Removal API identifies and removes unwanted shadows from photographs.

The AI Try-On API provides virtual garment fitting by overlaying apparel onto target models. Utilizing advanced computer vision, it creates realistic clothing simulations that account for body contours, fabric drape, and model poses.

The AI Model API generates fashion imagery by replacing models while preserving original outfits. It supports both predefined catalogs and custom model/scene uploads.

Human Matting extracts human figures from images using advanced segmentation algorithms to isolate subjects from backgrounds for use in content creation and virtual try-ons.

Clothing Matting isolates specific garments from images, enabling separate representation for e-commerce and virtual fitting applications.

Enables face swapping in videos by transferring facial features and expressions for dynamic and realistic results.

Improved video clarity for the Talking Face API while maintaining an output bitrate consistent with the source input.

The Magic Erase function removes specified objects or areas from an image based on user-defined coordinates.

Extracts portraits from images with professional matting effects.

Composites an extracted portrait into a new target image.

Extracts subject portraits from videos with high-precision matting.

Composites extracted video portraits into new video backgrounds.