Whispp raises €5M for on-device voice reconstruction
Whispp is using on-device audio AI to reconstruct speech in real time, combining an accessibility use case with a broader edge-AI distribution strategy.
Voice AI becomes more valuable when it can improve communication in real time without sending every interaction to the cloud.
What happened
Leiden-based Whispp raised €5M in follow-on funding led by LUMO Labs, alongside strategic angels and non-dilutive EIC Accelerator support.
Its audio-to-audio AI reconstructs speech in real time on-device, including turning whispered speech into a clearer natural-sounding voice.
The company is targeting people with voice impairments and speech disorders while also pursuing integrations with mobile and PC manufacturers.
Why it matters
This is more differentiated than a generic voice assistant.
Whispp is reconstructing characteristics of speech rather than simply transcribing words or removing background noise. Running the system on-device can also improve latency and privacy, both of which matter when voice processing is continuous and highly personal.
The OEM strategy is another important signal. Distribution through devices people already use could be more scalable than relying only on a standalone accessibility app.
The bigger picture
Edge AI is becoming more important where speed, privacy and continuous processing matter.
Voice is a strong example because even small delays can damage the experience. Whispp’s round shows how specialised on-device models can create value in areas where cloud-first AI is not always the best fit.