CEREPROC LTD

Their core work

CereProc is a speech technology company based in Edinburgh, UK, specializing in text-to-speech synthesis and the creation of natural-sounding synthetic voices. They build voice generation systems used in virtual assistants, conversational agents, and broadcast media. In their H2020 projects, they contributed speech processing and voice synthesis expertise — first to a consortium building virtual agents with natural language understanding, then to a project creating automated digital community radio platforms. Their core commercial product is a text-to-speech engine capable of producing highly natural, customizable synthetic voices, including for accessibility use cases such as replacing lost human voices.

What they specialise in

How they've shifted over time

CereProc's two H2020 projects show a progression from core AI research — building virtual agents with linguistic and social understanding (RIA scheme, 2015–2017) — toward applied digital media infrastructure, specifically scalable community radio platforms (IA scheme, 2018–2020). This shift from research-oriented conversational AI to practical deployment in media broadcasting reflects a maturation toward real-world product integration. The move to an Innovation Action in the second project suggests increasing confidence in deploying their technology in production environments beyond laboratory virtual agent prototypes.

CereProc appears to be moving from foundational speech AI research toward applied deployment in media and content production, making them a relevant partner for projects where voice interfaces or automated audio generation meet real end-user contexts.

How they like to work

CereProc has never led an H2020 project, always joining as a specialist participant — the pattern of a company that brings a specific, deployable technology to consortia rather than shaping the overall research agenda. Across just two projects they worked with 16 different partners in 8 countries, indicating they are actively sought by diverse consortium builders and comfortable operating in large, international teams. This breadth of partners per project suggests they are regarded as a valued specialist rather than a generic digital contractor.

16 unique consortium partners across 8 countries from just 2 projects, suggesting CereProc is consistently embedded in large, diverse European consortia. No clear geographic concentration beyond their UK base, though post-Brexit status may affect future eligibility for Horizon Europe participation.

What sets them apart

CereProc is a rare commercial SME in EU research that contributes production-grade speech synthesis technology — not academic prototypes — to research consortia. Their Edinburgh base and long-standing focus on highly natural synthetic voices, including the well-known use case of creating replacement voices for people who have lost the ability to speak, gives them a distinctive and socially resonant profile. For consortium builders in ICT, health-tech, or media, they represent a direct bridge between speech AI research and a deployable commercial TTS product with proven real-world deployment.

Highlights from their portfolio

• ARIA-VALUSPA
  CereProc's largest funded project (EUR 322,651), targeting virtual agents with both linguistic and social understanding — the core application space for their speech synthesis technology and a strong signal of their AI research credentials.
• GrassrootWavelengths
  Demonstrates CereProc's ability to pivot their voice technology into media broadcasting infrastructure, expanding their application reach beyond virtual assistants into scalable community radio platforms.