Duix: Silicon-Based Digital Human SDK 123

Silicon Intelligencehas open-sourced DUIX, a 2D photorealistic, AIGC real-time rendering digital human model! DUIX (Dialogue User Interface System) is an AI digital human intelligent interaction platform built by Silicon Intelligence. By open-sourcing digital human interaction capabilities, developers can independently integrate various large models, automatic speech recognition (ASR), and text-to-speech (TTS) capabilities to achieve real-time digital human interaction, and deploy with one click on multiple terminals including Android and iOS, allowing every developer to easily create intelligent, personalized digital human Agents and apply them across all industries.

Developers can independently integrate various large models, automatic speech recognition (ASR), and text-to-speech (TTS) capabilities to achieve real-time digital human interaction

It can be deployed with one click on multiple terminals including Android and iOS, and DUIX also provides 14 digital human templates

Supports low-cost and rapid deployment on iOS, Android, or large screens, and runs smoothly on tablets and in-vehicle systems

Supports smooth visuals at 50 frames per second, with video generation efficiency exceeding 1:0.5, meeting the high real-time requirements of scenarios such as live streaming

Accurately simulates movements, speaking lip shapes, and micro-expressions

DUIX Applicable Scenarios:

Low deployment cost: No need for the customer to provide a dedicated technical team to cooperate, supporting low-cost and rapid deployment across multiple terminals and large screens.

Low network dependency: Suitable for virtual assistant self-service scenarios such as subway stations, banks, and government services.

Diverse functions: Can meet the diverse needs of multiple industries including video, media, customer service, finance, and broadcasting, based on customer requirements.

DUIX Core Features:

Low deployment cost: No need for the customer to provide a dedicated technical team to cooperate, supporting low-cost and rapid deployment across multiple terminals and large screens.

Low network dependency: Suitable for virtual assistant self-service scenarios such as subway stations, banks, and government services.

Diverse functions: Can meet the diverse needs of multiple industries including video, media, customer service, finance, and broadcasting, based on customer requirements.

Real-time video communication: Users can have real-time conversations with digital humans.

Personalized customization: Users can choose their favorite appearance and voice based on personal preferences to create a unique interactive experience.

Multi-role play: Digital humans can become your companion, lover, or "encyclopedia" to meet your needs.

Emotional connection: Through interaction, users build emotional bonds with digital humans and enjoy the pleasure of companionship.

Moments: The digital human's "Moments" showcase their virtual life, allowing users to gain a deeper understanding of their personalities and interests.

DUIX Frequently Asked Questions:

1. Do you support digital human customization? Is it based on photos or videos?

Answer: Yes, we support digital human customization, and it is based on videos. You can customize an exclusive digital human for use in the SDK. However, customized digital humans are a paid service, and you can contact our customer service email.

2. How to customize the appearance?

Answer: Customizing a Silicon digital human requires filming a 3-5 minute video of a real person on camera delivering a script. The specific on-camera posture and script content can be determined based on the actual usage scenario. For example, if it is used to produce legal consultation-related videos, you can choose to appear in formal attire and deliver legal-related content, so that the overall scenario is more consistent.

3. How is the customization charged?

Answer: We provide digital human customization that pairs with the SDK open-source interface, priced at 9,800 yuan per set, including appearance and voice. For more customization needs, you can contact our customer service email.

4. How to update the digital human appearance?

Answer: At present, the open-source offering mainly provides public models. If you have specific needs, customization is available.

5. Is there an API interface for appearance cloning?

Answer: The training service currently only supports deployment on our internal servers for invocation.

6. Do the broadcast wav files support streaming data?

Answer: Streaming driver is currently being modified and optimized, and is temporarily not supported.

7. Are there callback methods for the start and end of broadcasting?

Answer: For the callback methods for the start and end of broadcasting, please refer to the SDK documentation.

8. Are there API controls for digital human movements?

Answer: Digital human movement APIs are temporarily not supported.

DUIX Commercial Case Showcase:

https://apps.apple.com/us/app/duix-your-ai-companion/id6451088879