DeCloakVision Avatar is a pioneering human-computer interaction solution designed to create highly intelligent and emotionally resonant digital personas. The system seamlessly integrates advanced Large Language Models (LLM), Speech-to-Text (STT), and Text-to-Speech (TTS) technologies, enabling users to engage in natural, real-time voice conversations via microphone. With precision-engineered visual rendering, the avatar displays lifelike body movements and nuanced facial expressions, significantly enhancing the sense of immersion and rapport. Product Design & Features： The core strength of the system lies in its Retrieval-Augmented Generation (RAG) capabilities, allowing the avatar to master domain-specific knowledge or real-time data to provide accurate, context-aware responses. This system demonstrates exceptional versatility across various sectors; it was prominently featured at the 2026 Yushan Forum and the international forum on “Local Democracy: Urban Innovative Governance under Global Collaboration,” showcasing its potential for digital consultation in democratic governance and urban innovation. Upholding DeCloak’s commitment to privacy, the system ensures that all

DeCloak Intelligences Co.  Launches FIDO2-Compliant Authentication Server to Enable Passwordless Public-Key-Based Verification Traditional passwords are vulnerable to theft, leading to data breaches and identity fraud. FIDO2 technology addresses these issues by introducing a public-private key mechanism for local authentication. The private key never leaves the user’s device, effectively eliminating the risk of credential leakage from servers and significantly enhancing account security. DeCloak Intelligences Co.’s FIDO2 server has been rigorously certified by the FIDO Alliance, ensuring full compliance with international standards and seamless interoperability. It empowers organizations to implement high-security, low-maintenance authentication infrastructures. The system supports various FIDO authentication devices, including security keys, mobile biometrics (such as fingerprint and facial recognition), and integrates easily with existing IT environments. In the face of increasing cybersecurity threats, DeCloak Intelligences Co. positions FIDO2-based strong authentication at the heart of digital transformation and enterprise security. DeCloak Intelligences Co. has officially achieved FIDO2® Server Certification, successfully

DeCloakBrain is a groundbreaking, privacy-aware, multimodal intelligent robotic system featuring advanced AI technologies like Privacy-Secure Perception, Super-Visual Synthesis, Cognitive Inspiration, Spatial Intelligence and Robot/Machine Collaborative Learning. Utilizing multimodal AI models with federated learning, differential privacy and homomorphic encryption, it ensures the highest privacy and quantum-safe security. DeCloakBrain’s plug-and-play design seamlessly integrates with robots/machines via hardware interfaces, providing omnidirectional spatial visualization and panoramic spatial perception. Standardized front-end and back-end configurations enable efficient AI deployment/management on edge and private cloud/server platforms. Its innovative Privacy Visual Language Model ensures efficient visual perception processing and secure de-identification, safeguarding sensitive information and human privacy rights. Design： DeCloakBrain prioritizes human privacy and data security while delivering efficient robotic intelligence with panoramic spatial perception via super-visual synthesis. The well-designed, privacy-secure AI enables operators to track specific activities, such as access control and security threat detection, without compromising privacy. Authorized personnel can prompt robots/machines via voice input, enhancing

DeCloakVision is an innovative privacy-enhancing surveillance system that employs advanced techniques to protect individuals’ privacy while enabling the targeted observation and tracking of specific activities. The system utilizes decentralized AI models, leveraging well-designed differential privacy and homomorphic encryption technologies to facilitate secure and anonymous tracking of individuals without compromising their identities or the privacy of the general public. DeCloakVision’s cutting-edge multi-modal deep neural network (DNN) model ensures high efficiency in human image processing tasks and accuracy in recognition tasks, while providing differentially private de-identification and quantum safe encryption to safeguard image data. The system seamlessly integrates with existing surveillance infrastructure and offers a robust solution to address the challenges of mass surveillance while upholding individual privacy rights. Privacy-Preserving Al Operating ModeI： Features of DeCloakVision：

Face recognition has manifested itself as a cutting-edge solution for various industries as it secures access control, boosts customer experience, provides strong authentication, promotes goods and services, and speeds up financial operation. However, those profits often come at the expense of users, such as privacy and even safety. This constrains the development of capital market for facial recognition. DeCloakFace provides privacy protected identity authentication that image processing is done at local devices only so that no sensitive personal facial photos will be uploaded to cloud to avoid data leakage. The innovation solution has brought a great deal of attention from industries including: finance, healthcare, cryptocurrency and E-signing platforms, where ensuring privacy while collecting data is crucial.

For the purpose of ensuring data security and confidentiality, DeCloak has developed a solution: Enterprise File Secure Exchange Platform, which combines multi-factor authentication, such as facial recognition, email verification and etc., to ensure the qualifications of file viewers and protect data/files confidentiality within the company.