CandyVoice is a French deep-tech company developing a modular portfolio of real-time AI and DSP technologies for voice processing in embedded, edge and distributed computing environments.
Our technologies include:
AI-based noise filtering and speech enhancement for extreme acoustic environments
Low-bitrate voice coding and packet-loss resilience (PLC) for robust communications
Real-time voice anonymisation for privacy-preserving voice processing
Real-time voice deepfake detection for trustworthy communications and content authenticity
Real-time voice imitation
CandyVoice technologies are designed for ultra-low latency, low computational footprint and local processing. They can be integrated into embedded devices, edge nodes, gateways, servers and distributed infrastructures, enabling voice AI services to operate close to the data source while reducing cloud dependency, bandwidth requirements and exposure of sensitive voice data.
For HORIZON-CL4-2027-04-DATA-08, CandyVoice can contribute resource-efficient real-time AI voice components to distributed and decentralised computing environments, including resilient communications, edge processing and privacy-preserving voice services.
For HORIZON-CL4-2027-04-DATA-03, CandyVoice can contribute deployable AI voice components across the Edge-Cloud Continuum, supporting low-latency processing, distributed execution, interoperability and the integration of AI services from constrained edge devices to infrastructure-level platforms.
CandyVoice technologies are mature and deployment-oriented, with solutions reaching TRL 6–8 depending on the component and use case. CandyVoice received a 2026 International Critical Communications Award (ICCA) – “Best Use of Advanced Technology” for its Mission-Critical AI Voice Stack.
We are seeking consortium coordinators, research organisations, edge/cloud platform providers, system integrators and industrial partners preparing proposals for DATA-08 and DATA-03, where CandyVoice can contribute specialised real-time voice AI components, integration expertise and validation capabilitie.
Cluster 4: Digital Technologies
CEO / CTO
CandyVoice
Business Development
CandyVoice
Project cooperation
HORIZON-CL4-2027-04-DATA-08 – Edge AI & Real-Time Voice Technology Partner for Distributed Computing
CandyVoice seeks to join a DATA-08 consortium as a technology partner, contributing ultra-low-latency, resource-efficient AI voice components for distributed edge environments, resilient communications and privacy-preserving voice processing.
CEO / CTO at CandyVoice
Strasbourg, France
Project cooperation
HORIZON-CL4-2027-04-DATA-03 – Real-Time Voice AI Technology Partner for the Edge-Cloud Continuum
CandyVoice seeks to join a DATA-03 consortium as a technology partner, contributing ultra-low-latency, resource-efficient Voice AI components deployable across embedded devices, edge nodes, gateways and cloud-edge infrastructures.
CEO / CTO at CandyVoice
Strasbourg, France
Expertise
Real-Time, Resource-Efficient Voice AI for Edge & Distributed Computing
CandyVoice provides ultra-low-latency, resource-efficient Voice AI expertise for embedded, edge and distributed computing: speech enhancement, resilient voice communications, privacy-preserving anonymisation and deepfake detection.
CEO / CTO at CandyVoice
Strasbourg, France
Product
FrameFix – Real-Time Audio Frame Reconstruction for Resilient Digital Communications
Real-time technology reconstructing missing speech frames under severe packet loss. Codec- and transport-independent, FrameFix runs at the edge to maintain voice intelligibility and improve speech processing across degraded digital networks.
CEO / CTO at CandyVoice
Strasbourg, France
Product
VoxScan – Real-Time Voice Deepfake Detection for Trustworthy AI & Digital Communications
Real-time technology detecting synthetic, cloned and manipulated voices without a reference voice model. Language-independent and GPU-free, VoxScan runs at the edge and can strengthen trust and authenticity in European AI and digital ecosystems.
CEO / CTO at CandyVoice
Strasbourg, France
Product
NoizzOff – Real-Time Edge AI Noise Reduction for Robust Voice Processing
Embedded Edge AI technology restoring intelligible speech in extreme noise with >32 dB noise reduction and only 16 ms latency. NoizzOff runs locally on low-power processors and improves voice quality for communication and AI applications.
CEO / CTO at CandyVoice
Strasbourg, France
Product
VoxShield – Real-Time Voice Anonymisation for Privacy-Preserving AI & Digital Communications
Real-time technology removing biometric voice identity while preserving speech intelligibility and ASR usability. VoxShield runs at the edge, enabling privacy-preserving voice processing, AI training and secure digital communications.
CEO / CTO at CandyVoice
Strasbourg, France