Soul App Redefines Human-AI Interaction with Full-Duplex Technology at WAIC 2025

SHANGHAI, July 30, 2025 /PRNewswire/ -- At the 2025 World Artificial Intelligence Conference (WAIC) on July 26th, Soul App unveiled its AI-powered social solutions. In a live demonstration of its "Audio Partyroom", the platform showcased its upgraded end-to-end full-duplex voice large language model (LLM), where an AI host seamlessly moderated group voice chats. The session highlighted how AI reshapes content and social dynamics — delivering smarter, more authentic, and deeply immersive emotional experiences.

"In this era where technology redefines human connection, AI-powered social networks aren't just the future—they're the present", declared Tao Ming, Chief Technology Officer of Soul App. "We're architecting a revolutionary paradigm where AI beings and humans don't merely interact but thrive together. Through intelligent connection mapping, contextual understanding, and experience augmentation, Soul is cultivating a new dimension of digital belonging—one that delivers genuine joy and transforms how we perceive community".

Toward More Natural AI Interaction: Full-Duplex Voice & Video for Genuine Emotional Connection

As technological advancements mature in 2025, the "AI application explosion" has become an industry consensus. Amid AI's sweeping transformation across sectors, a question remains: What concrete value does "AI+" bring to individuals?

Representing the forefront of AI social exploration, Soul App presented its vision at WAIC 2025, demonstrating social solutions that redefine AI's role in human connections.

At the exhibition themed "Soul's AI Social Practice", the platform highlighted how its full-duplex audio/video capabilities enable hyper-realistic and deeply immersive social experiences. Soul's AI serves dual roles: as a caring companion that analyzes users' social challenges to provide emotional support, and as an efficient connector that identifies mutual interests between users to foster genuine friendships.

The event marked the debut of Soul's self-developed, end-to-end full-duplex voice model. Recently upgraded, this innovative model redefines the interaction paradigm by eliminating traditional reliance on Voice Activity Detection (VAD) mechanism and latency control logic. It breaks away from the industry's prevalent turn-taking dialogue mode, empowering AI to autonomously determine conversation rhythms.

Social interaction involves the exchange of both emotional and informational value. By endowing AI with autonomy and warmth, Soul enables truly reciprocal social engagement and emotional symbiosis between humans and AI, moving beyond the transactional turn-based dialogues of the past.

The full-duplex voice model will soon enter beta testing before being implemented across various scenarios, including virtual human real-time calls and AI-assisted matching in both one-on-one and group settings. At WAIC, attendees experienced this technology firsthand through an AI host managing group voice chats on Soul's platform.

In multi-user voice interactions, the AI host demonstrates human-like moderation capabilities, naturally engaging participants, energizing conversations, and encouraging discussion—significantly enhancing engagement and experience quality. "Applying full-duplex voice technology to group scenarios represents a crucial step in integrating AI into social networks, enhancing real connections, and enabling diverse experiences", said Tao Ming, Soul's CTO.

Recognizing users' inherent preference for face-to-face interaction, Soul is also advancing multimodal capabilities. The exhibition featured the company's latest progress in full-duplex real-time video call, combining audio technology with real-time video generation to let users see "the AI's visual presence and digital environment," creating more dimensional digital personas that deepen relational bonds. This builds on Soul's recent achievement of having its real-time portrait video generation research accepted by CVPR 2025, a top-tier international academic conference.

Soul Reimagines Social Networking: Building an AI-Powered Playground for Chinese Gen Z

Since its launch in 2016, Soul has been at the forefront of reinventing social discovery, connection, and relationship-building through AI. The platform empowers users to express themselves in a pressure-free environment while forming diverse social networks based on shared interests—exemplified by its Lingxi Engine, the foundation of its current relationship model.

Soul's AI journey accelerated in 2020 with systematic AIGC R&D, yielding cutting-edge capabilities in intelligent dialogue, voice technology, and 3D digital humans. The 2023 introduction of its proprietary Soul X model marked a turning point, followed by breakthroughs including voice generation, voice call, and music generation models. By 2024, Soul had consolidated these into a comprehensive multimodal end-to-end AI system supporting text/voice conversations, multilingual understanding, and hyper-realistic digital humans.

The platform's competitive edge stems from three pillars: in-house R&D, unique public social data assets, and rapid application iteration. As a social networking platform, Soul continuously refines its "Emotional Intelligence Model" to precisely identify user needs and deliver lasting emotional value through AI.

Soul's deep market insights—drawn from its user community—directly inform product development. Notably, 18% of new posts on Soul Square address real social challenges (making friends, starting conversations, workplace relationships). These authentic user stories and the emotional connections they spark form a valuable "social behavior database".

Building on these insights, Soul is deepening AI-community integration by developing content recommendations and AI-assisted creation tools that lower expression barriers, while also cultivating an ecosystem of "Virtual Influencers" capable of delivering warm, nuanced emotional support—all working in concert to create authentic senses of emotional closeness within digital spaces.

At its core, the platform operates on a dual approach: facilitating human connections through AI assistance while simultaneously fostering meaningful "AI-as-companion" relationships". True AI integration isn't just about technology—it's about weaving genuine emotional connections into the very fabric of users' social networks", explained Tao Ming, Soul's CTO. "As AI becomes deeply embedded in every aspect of modern interaction—from socializing to digital lifestyles and consumer behavior—it's giving rise to what we call emotional-value commerce. This evolution isn't merely creating new market opportunities; it's fundamentally reimagining the paradigms of social products".

 

** The press release content is from PR Newswire. Bastille Post is not involved in its creation. **

Soul App Redefines Human-AI Interaction with Full-Duplex Technology at WAIC 2025

LAS VEGAS, Jan. 17, 2026 /PRNewswire/ -- At CES 2026, Flowtica was not the loudest presence on the show floor. Over the course of the week, however, the Singapore-based AI company became one of the most closely watched, drawing sustained attention from media, investors, and industry observers.

That attention has now translated into availability. Following its appearance as a CES 2026 Innovation Award Honoree, Flowtica has officially opened sales of Flowtica Scribe, its AI recording pen that quietly stood out in a field dominated by phone-dependent and screen-heavy solutions.

Flowtica Scribe did not make its first public appearance at CES. The product launched on Kickstarter several months earlier, where it attracted thousands of users and established early validation. CES marked a different stage for the company. The focus shifted away from introduction and toward closer examination, with many observers asking whether the idea could hold up beyond early enthusiasm.

Throughout the exhibition, Flowtica's booth at Eureka Park became a steady meeting point for investors and journalists. Conversations rarely centered on specifications or feature lists. Instead, discussions focused on a more fundamental question: whether AI recording tools are beginning to converge on forms that genuinely fit into professional life.

Flowtica's answer was visible in its choice of form. By embedding AI into a pen, an object already accepted in boardrooms, consultations, investor meetings, and sales conversations, the company avoided many of the social and practical frictions that continue to limit phone-based recorders. The result was a device that felt less like a new category of gadget and more like a natural extension of existing behavior.

This restraint is deliberate. Flowtica Scribe is designed to remain unobtrusive, with no screen and no demand for user attention during conversations. Recording, organization, and interpretation take place quietly in the background, allowing users to remain focused on the discussion itself. Among on-site observers, this approach positioned Flowtica as one of the more pragmatic entrants in an increasingly crowded AI recording market.

Hardware, however, represents only part of Flowtica's differentiation. The company places equal emphasis on what happens after recording ends. Its AI functions as a continuously evolving insight system that can operate autonomously while remaining responsive to clear user guidance when needed. Over time, it adapts to individual working rhythms and priorities, shifting the emphasis from capturing everything to identifying what truly matters.

Rather than simply storing conversations, Flowtica helps determine which moments should be retained, revisited, and translated into next actions. Key insights are designed to integrate naturally with existing productivity tools, including calendars and task systems, allowing them to fit into established workflows rather than compete with them.

As interest in Flowtica Scribe built during CES, one question surfaced repeatedly. When would the product be available beyond the show floor?

With sales now officially open through Flowtica's website, the company moves from exhibition attention to real-world deployment. For Flowtica, this moment appears less like a conclusion and more like a transition, from being observed to being used.

As the AI recording market continues to mature, competition is shifting away from novelty and early adoption toward durability. The defining question is no longer whether a system can record conversations, but whether it can remain useful over time.

Flowtica enters this next phase with an approach that is measured and quietly confident, placing its bet not on spectacle, but on long-term relevance.

** The press release content is from PR Newswire. Bastille Post is not involved in its creation. **

The AI Recording Pen That Turned Heads at CES Is Finally for Sale