The “Convo AI & RTE 2025 Real-Time Internet Conference,” jointly hosted by Agora and the RTE Developer Community, recently concluded in Beijing. The conference highlighted how conversational AI (Convo AI) is driving a new wave of innovation in real-time engagement (RTE), extending the boundaries of interaction from connections between people to a new stage of interaction between people and AI.

As a leading company specializing in computer graphics and artificial intelligence, FaceUnity was invited to attend and showcase its innovations in AI digital humans and intelligent imaging effects for real-time engagement. By creating more intuitive, immersive, and diverse experiences, it continues to advance the evolution of human–computer interaction.

Digital Humans:

Creating a New Model of Emotional Interaction Through Realism and Immersion

For FaceUnity, digital humans are more than a bridge between virtual and physical worlds: they are a medium for the next generation of emotional interaction. The core of this interaction is highly visual, humanlike presentation combined with perceptive and expressive intelligence comparable to that of a real person, creating a new space for emotional interaction that feels more realistic and immersive.

Guided by this vision, FaceUnity builds on its own research and development to create a full-stack digital human technology system. It focuses on breakthroughs in driving, interaction, rendering, and other core components to enable genuinely humanlike interaction with digital humans.

In visual performance, FaceUnity uses internationally leading robust driving and real-time animation technologies to significantly improve the stability and precision of digital human animation. It also enhances lip-sync accuracy and the computational efficiency of real-time animation, bringing body movements, facial expressions, hairstyles, clothing, and other details from basic visibility to lifelike expressiveness.

In interaction, FaceUnity’s proprietary multimodal natural interaction technology lets users engage through a monocular camera, speech, text, and other channels. Combined with conversational AI’s deep semantic understanding and content generation, a digital human can interpret a user’s expressions and understand their intent. It responds to dialogue within seconds and synchronizes lip movement with audio in milliseconds, making each interaction as natural, fluid, and emotionally engaging as a conversation with a person.

In rendering and deployment, FaceUnity’s proprietary lightweight real-time rendering engine surpasses mainstream commercial engines in feature completeness, ease of integration, and installation package size. It enables digital humans to run smoothly on phones, large displays, VR devices, and other terminals, offering low-latency, immersive visual interaction wherever users are.

Notably, drawing on its full-stack expertise, FaceUnity pioneered a real-time monocular 3D motion capture system for mobile devices. It was the first worldwide to capture body, finger, and facial movements at 30 FPS using only a smartphone RGB camera, with joint error below 5 mm. This sets a technical benchmark for high-precision visual interaction with digital humans on lightweight devices and lays the groundwork for widespread adoption on mobile terminals.

AR Video Effects SDK

Opening the Global Social Entertainment Market Through Broad Compatibility and Deep Optimization

In intelligent imaging, FaceUnity’s AR Video Effects SDK is becoming a powerful tool for developers worldwide to build differentiated products, thanks to its exceptional cross-platform compatibility and deep localization.

When virtual social platforms, live streaming, and short-form video first grew rapidly, interactive features such as video effects, intelligent beauty enhancement, and creative stickers drew considerable attention. To meet this demand, FaceUnity launched its AR Video Effects SDK based on advanced AI algorithms for faces, bodies, gestures, and segmentation. It brings together intelligent beauty and makeup, AR stickers (including 2D, 3D, and AR mesh), portrait segmentation, chroma keying, gesture and expression recognition, and other features to improve user interaction on short-form video, live-streaming, and audio and video calling platforms.

As effects become more varied and use cases more diverse, however, device performance and platform compatibility have become critical. In overseas markets such as Southeast Asia and the Middle East, platforms and developers must deliver high-quality visuals while also keeping complex effects smooth and bringing them into commercial use.

In response, FaceUnity was among the first to make its AR Video Effects SDK fully compatible with major platforms including HarmonyOS NEXT, iOS, Android, Windows, macOS, Unity, Flutter, and Electron. It also optimized power consumption on entry-level and mid-range devices, striking a strong balance between visual quality and smooth performance.

To meet the differing needs of global markets, the AR Video Effects SDK is also tailored to different skin tones, aesthetic preferences, and cultural backgrounds, extending a high-quality interactive experience seamlessly to a wider global audience.

To date, FaceUnity’s AI digital human and intelligent imaging effects products and solutions have been deployed in 22 vertical sectors, serving more than 2,000 enterprises worldwide across over 200 countries and regions.

In digital human interactions, FaceUnity’s digital humans act as helpful “smart assistants” and “travel companions” on intelligent devices, providing personalized services with emotional warmth. The company has established close partnerships with several leading smartphone and automotive manufacturers to move the user experience from meeting functional needs toward building emotional connections.

In video interaction, the AR Video Effects SDK offers varied creative features that make social and entertainment experiences more engaging and immersive. It has been adopted by well-known overseas products such as India’s short-form video platform Josh and the U.S. social app MeetMe.

As the real-time engagement industry continues to develop, a more intelligent, immersive, and human-centered digital world is coming into view. With digital humans and imaging effects as its core technologies, FaceUnity will help shape a future of human–computer interaction with deeper integration, emotional connection, and limitless use cases.