[Automatic translation follows] Advanced DSP & AI voice technologies for secure communications and immersive virtual worlds.

CandyVoice is a deep-tech company specializing in real-time voice technologies, combining advanced DSP and artificial intelligence.
We develop innovative voice software bricks for dual use, intended for secure strategic communications as well as virtual worlds and immersive environments.

Our core expertise concerns real-time processing of the human voice: expressive vocal imitation, irreversible anonymization, protection of voice identity, noise reduction, recovery of lost frames and detection of synthetic or manipulated voices. These technologies make voice more immersive, safer and more reliable, without resorting to generic synthetic voices.

CandyVoice solutions are designed privacy-by-design, compliant with European requirements (GDPR, AI Act), and optimized for constrained environments: edge computing, embedded systems, telecom infrastructures and XR platforms. Our hybrid DSP + AI approach guarantees ultra-low latency, reduced consumption and easy integration on general or specialized processors (CPU, ARM, NPU).

CandyVoice positions itself as a technological supplier of voice bricks and R&D partner in European and industrial projects, particularly in the fields of virtual worlds, human-centered AI, telecommunications, security and critical communications.

Noise Reduction, Noise Concealment, Packet Loss Concealment, PLC, Radio communication, Telecom, C4ISR, GDPR, Noise Filter, Voice Deepfake Detection, Voice Anonymization, AI, IA, DSP, Digital Voice Processing, Digital Voice Processing, and Digital Signal Processing