Loquela — Android Voice Keyboard

I built Loquela, an Android voice keyboard with on-device and cloud speech recognition, and its supporting web and backend components.

My work includes the native Kotlin app, model selection and downloads, keyboard prediction and autocorrect, and Google Play entitlement checks. The stack includes sherpa-onnx, speech-recognition models, Next.js and Firebase.

The project brings together model behavior, device resources, privacy, usability and paid access. My role spans implementation, testing and iteration across those pieces.

Public product site: https://www.getloquela.com/

The cover is a project-summary graphic, not a product screenshot.