Flowent

• 

  Interactive stories based on user level, interests, and proficiency, including audio playback and key vocabulary.

• 

  Scenario selection library allowing users to choose from diverse real-world contexts and creating a custom scenarios.

• 

  Immersive roleplay interface with a dynamic sidebar checklist, allowing users to track completed tasks and conversation goals.

• 

  Post-lesson analytics dashboard providing instant feedback on completion time, accuracy scores, and task summaries.

Inspiration

We built Flowent after repeatedly experiencing the same problem as language learners: speaking is the bottleneck. After B1/B2, students don’t get enough speaking time in class, and institutions can’t scale it.

What we built

Flowent enables language programs to assign structured speaking tasks that students complete asynchronously using real-time voice AI. Students practice speaking; programs get measurable insights into progress.

How it works

We use real-time speech-to-text, low-latency conversational AI, and scenario-based prompts to simulate realistic speaking situations. Every interaction is tracked and evaluated.

Why Gemini

Gemini powers fast, context-aware conversation generation and multilingual handling, allowing us to simulate real speaking scenarios with low latency and high reliability.

Impact

Flowent helps programs give every student more speaking time—without increasing teacher workload—while making speaking progress measurable instead of subjective.

Built With

• firebase-firestore
• google-gemini-api
• node.js-(express)
• react-native-(expo)
• supabase-(postgresql)
• tts
• versal