Welcome toAkoli Speech

Helping people living with speech that may be difficult to understand — with personalized speech recognition that learns to better recognise how you speak.

Akoli Speech helps people living with different speech to create and use their own personalized automatic speech recognition (ASR) model — that better recognises individual speech patterns.

A person records a set of speech samples in their own time, and the app learns from them to build a personalized speech recognition model. The goal is to more accurately caption the words you say.

The app builds on research from Google's Project Euphonia.

Akoli Speech is developed at the Centre for Digital Language Inclusion (CDLI) at UCL's Global Disability Innovation Hub (GDI Hub), in collaboration with Talking Tips Africa and Senses Hub.

The development of Akoli Speech has been supported by Google.org and the UK International Development Funded AT2030 programme that tests ‘what works’ to improve access to life-changing Assistive Technology (AT) for all. Global Law firm Hogan Lovells provided pro bono legal advice.

Launching soon.Check back here for updates.