Back to projects

Flowent — Real-Time Voice Agent for Language Immersion

A production voice agent: streaming bidirectional audio, sub-second turn latency, and multi-turn speech interaction, grounded in cognitive-science immersion methodology.

Live in production on the App Store and Google Play, with sub-second turn-taking and streaming TTS.

Flowent is a language-immersion app built around a real-time voice agent, now live on the App Store and Google Play. The research-relevant substance is the live conversational layer: streaming bidirectional audio, sub-second turn latency, and natural multi-turn speech interaction — expressive and multimodal speech running in production, in front of real users.

What it does

  • Real-time, full-duplex voice conversation with sub-second turn-taking
  • Streaming TTS so the agent starts speaking before the full response is generated
  • Cognitive-science-grounded immersion methodology driving the interaction design
  • AI-powered personalization of pacing and difficulty

Features

  • AI chat integrated across the whole app — a tutor available in lessons, scenarios, stories, and review, not a separate mode
  • SRS engine (SM-2) — every word and phrase tracked and resurfaced right before it would be forgotten
  • Stories for reading and listening — level-appropriate stories with audio, including custom stories generated on demand
  • Games — short, fast practice loops for low-friction daily review