Stepping the Future of Social+AI Why You Should Watch Soul App Shine

The tech scene is buzzing with excitement as the 8th International Conference on Computer Graphics and Virtual Reality (ICCGV 2025) recently announced its paper acceptances. A groundbreaking submission from Soul App, a popular Chinese social networking platform, “LLM Gesticulator: Leveraging Large Language Models for Scalable and Controllable Co-Speech Gesture Synthesis,” is making waves at the event.

ICCGV is a hotspot for innovation in computer graphics and virtual reality (VR), exploring cutting-edge topics like VR, augmented reality (AR), mixed reality (MR), human-computer interaction, and advanced immersive experiences. With artificial intelligence (AI) propelling VR into an exciting new phase, the conference attracts a mix of academic powerhouses, industry leaders, and creative pioneers to share ideas and foster collaboration. Amid this tech extravaganza, Soul App’s inclusion underscores its growing influence as a disruptor in AI-driven social technologies.

Soul’s paper highlights its innovative use of large language models (LLMs) to revolutionize multi-modal drives in virtual humans. By leveraging these advanced models, Soul has set new benchmarks for animation quality, relevance, rhythm alignment, and text matching. The paper’s acceptance marks a major milestone for Soul and showcases its commitment to pushing the boundaries of interactive technology for younger, tech-savvy users.

Since its launch in 2016, Soul has consistently championed the integration of AI and virtual environments. Unlike traditional social platforms, Soul aims to minimize social pressures by eliminating the need for real profile pictures. Instead, users can craft unique virtual avatars through its intuitive avatar creation system. This approach has resonated deeply with younger generations, fostering a more inclusive and creative online space.

Soul’s journey into cutting-edge tech began with the launch of its proprietary NAWA engine, a blend of AI, rendering, and image processing technologies. This breakthrough allowed users to design personalized 3D avatars and virtual scenes easily. By 2020, Soul took a bold leap into AI-generated content (AIGC), diving into intelligent conversation, voice technology, and 3D virtual humans. Fast-forward to 2024, and Soul has upgraded its AI capabilities to a multi-modal end-to-end model, enabling seamless interactions across text, voice, and visuals. These advancements offer users near-human emotional companionship, creating natural and engaging experiences.

Behind Soul is founder Zhang Lu, who has long focused on reducing social pressures by enhancing user experiences online. Today, Zhang Lu is doubling down on multimodal integration, aiming to enhance 3D virtual human interactions with AI-powered solutions.

The “LLM Gesticulator” paper is a testament to Zhang Lu’s strategy. It introduces a revolutionary approach, using LLMs as the backbone for driving 3D virtual humans. Unlike conventional models, this solution simultaneously supports text and audio inputs, producing animations that synchronize perfectly with speech rhythms and textual meanings. The model also employs automated data annotation using vision-based AI, paving the way for more efficient and accurate training processes.

Soul’s contributions to AI research have earned it significant recognition in academia and industry. Last year, the platform’s paper on multimodal emotion recognition was spotlighted at the ACM Multimedia Conference’s MRAC 24 workshop. By enhancing the accuracy and scalability of emotion recognition, Soul demonstrated its commitment to tackling complex challenges in AI.

Soul’s large language models and interaction solutions are already transforming the user experience. From the playful AI chatbot “AI GouDan” to immersive gaming agents and digital twins, these innovations are redefining social interactions. At GITEX GLOBAL 2024, one of the world’s premier IT exhibitions, Soul wowed attendees with its 3D virtual human solutions. Visitors could create digital avatars in real time and engage in lifelike, multi-modal interactions, blending the boundaries between physical and virtual realities.

Looking ahead, Zhang Lu plans to expand Soul’s 3D virtual human capabilities across diverse scenarios. Whether it’s virtual companionship, group chat parties, or next-gen gaming, the platform is gearing up to offer users richer, more immersive experiences. By merging cutting-edge AI with social dynamics, Soul empowers users to express themselves at ease, share authentic emotional connections, and interact in a new dimension.

Soul App is at the forefront in a world where technology is reshaping how we connect. It proves that the future of socializing is not just about communication but also about creating meaningful, emotionally resonant interactions. With pioneering research and innovative applications, Soul is turning the vision of seamless, immersive virtual experiences into reality.