Polyglot is a cross platform AI language practice application that runs as a desktop app and also offers a web version. It is built around conversational large language models and Azure based text to speech services, turning them into an interactive environment for speaking practice in multiple languages. Users can define custom AI personas, choose languages, and configure their own OpenAI and Azure keys so they retain control over which backends they use. The app supports speech recognition with quick keyboard shortcuts, allowing learners to hold down a key to speak and release it to submit for recognition and response. It includes translation features, dark mode, playback of the user’s own recorded speech, and word highlighting that tracks the progress of synthesized audio to make following along easier. Polyglot also integrates additional AI providers, supports configurable conversation scenarios, and lets users personalize avatars, making the experience more engaging and flexible.
Features
- Cross platform desktop app plus web version for language practice
- AI driven conversations using configurable OpenAI and Azure backends
- Azure based text to speech synthesis with per word highlighting
- Speech recognition with push to talk style keyboard shortcuts
- Customizable languages, personas, avatars, and conversation scenarios
- Built in translation, dark mode, and playback of user recordings