1. Basic Information
MiracleVision is Meitu's proprietary AI visual model and open platform, positioned as an "AI visual engine that understands aesthetics." The platform provides standardized APIs and console services to businesses and developers, focusing on capabilities such as text-to-image, image-to-image, intelligent photo editing, and video generation. The platform also delivers high-quality visual content for industries such as e-commerce, advertising, gaming, animation, and film. Leveraging Meitu's extensive research on aesthetic trends and extensive aesthetic data, MiracleVision has established an aesthetic training and evaluation system, with stable image output, consistency, and industrialized delivery as its core keywords.
2. Product Overview
MiracleVision provides services through both an online console and an API, forming a closed loop from material import, parameter configuration, generation, review, and distribution. The model is specifically optimized for character, fashion, product, scene, and visual design tasks, covering multi-stage production needs from concept sketches to commercial films. The platform provides controllable methods such as text prompts, reference image constraints, and style templates, and supports zooming in, redrawing, background changes, and composition extensions, balancing quality and efficiency. For video scenarios, the platform provides image-to-video and video-to-video capabilities for the rapid production of promotional videos, short videos, and animated content. The platform supports account-based resource management, call statistics, and key permission control to facilitate team governance and compliance audits.
3. Core Functions
1. Main functions
Generating text-based and image-based images, supporting high-quality images of people, products, and scenes
Image editing suite including cutout, beauty and makeup, HD enlargement, restoration and stylization
Image-to-video and raw video, enabling storyboard expansion, animation, and multi-scale export
Style templates and prompt engineering tools to improve consistency and batch production efficiency
Enterprise console and log auditing, supporting key management, call monitoring and usage reporting
2. Technical characteristics
Aesthetic data and evaluation system, taking into account both detail fidelity and style unity
Multi-task and multi-industry optimization, covering e-commerce, advertising, games, animation, and film and television
Controlled production and consistency strategies reduce deviations and lower rework costs
Standardized APIs and SDKs enable quick integration into existing systems and workflows
4. Pricing and Versions
MiracleVision offers both pay-as-you-go and subscription models, with differentiated benefits including call quotas, concurrency, and priority queues. Enterprise users can apply for higher concurrency and customized parameters, including tiered permissions and project isolation. Pricing, quotas, and value-added benefits are subject to dynamic adjustments based on operational strategies and version iterations. Billing, tax, and compliance terms may vary by region or channel, and are subject to the official console and service agreement.
5. Applicable Scenarios and Target Audience
Rapid production of product main images, model try-ons, and event KV for e-commerce and brand operations
Creative expansion, multi-size extension and landing material generation for advertising and marketing teams
From sketches to finished pieces of character, scene, and prop design for games and animation art
Storyboard expansion, poster material and stylized shot generation for film, television and short videos
Image and video generation API integration and automated production pipeline for developers and ISVs
6. Frequently Asked Questions
Q: What access methods and permission management does MiracleVision support?
A: It supports direct console connection and API/SDK access, provides key management, sub-account and call limit settings, and can isolate permissions and billing by project dimension.
Q: How to ensure the controllability of the model in human and e-commerce product scenarios?
A: We achieve consistency in composition and texture through reference image constraints, style templates, and parametric control, and provide post-processing tools such as magnification, redrawing, and background replacement for stable delivery.
Q: Does it support video-related capabilities?
A: It supports both image-generated video and raw video, which can be used for promotional videos, animations, and storyboard extensions. The output specifications and speed are related to the materials and concurrency, and the actual output is subject to the current capability description.
Q: How to choose billing and quota?
A: You can choose the call quota and concurrency level by volume or subscription. Enterprises can apply for higher concurrency and customized parameters as needed. Prices and benefits are subject to the console and agreement page.
Q: How to deal with compliance and data security?
A: The platform provides a compliance mechanism for data transmission and storage, supports log auditing and illegal content processing; retention and deletion policies are based on the privacy policy and terms of service and vary by region.
