FluencyPath

FluencyPath motivates streaks in improving your English fluency. Track your level with AI feedback and curated videos for vocabulary/communication to reach the next stage daily.

Comment

Inspiration

Many learners get stuck because they lack a structured roadmap. They consume content passively but don't know if they are actually getting better. FluencyPath was born out of the realisation that people need a guided, step-by-step path to maintain their practice streak. We wanted to build a "daily companion" that doesn't just provide resources, but actively tells you, through data and AI feedback, whether you are improving every single day, turning uncertainty into tangible progress.

What it does

FluencyPath provides a step-by-step guided journey to mastery, ensuring you stay motivated and never lose sight of your progress. It solves the uncertainty of self-study through:

  • Guided Listening: Curated video paths selected to sharpen your vocabulary and communication skills based on your specific level.
  • Active Speaking: a safe space to practice 250+ real-world interview questions with instant feedback on how you sound and how to improve.
  • Precision Writing: Daily practice with deep-dive analysis into your grammar and word choices to help you express yourself more clearly.
  • Progress Awareness: Know exactly where you stand with a dashboard that tracks your level and visualises your growth over time.
  • Daily Motivation: Stay consistent with a streak system and personal AI guidance that keeps you moving toward the next stage of fluency.

How we built it

FluencyPath was engineered to be fast, secure, and user-centric. We built the frontend with React and Tailwind CSS to create a focused, distraction-free "YouTube Matrix" interface. The backend runs on a Node.js server that securely acts as a bridge between the user and our AI engine.

For the core intelligence, we integrated the Google Gemini API to act as an on-demand tutor, analysing speech and text to provide CEFR-level feedback. Data and user progress are stored in Google Firestore, ensuring real-time syncing across devices. We also implemented custom-built visualisation tools to render progress charts directly in the browser, ensuring the app remains lightweight and responsive without needing heavy external libraries. Security was a priority, so we implemented robust password hashing and role-based access control to protect user data.

Challenges we ran into

Our journey began in Google AI Studio, but we hit immediate roadblocks. The embedded videos refused to play, and despite multiple attempts, the sandbox environment couldn't support the interactive media features we needed. This forced us to pivot: we exported the project to a local machine and continued development using the Gemini CLI Agent. https://ai.studio/apps/drive/1OogatYrdmrDVzKKBTR3RB9sewzjIjA-L

This transition introduced "The Secure Context Paradox." Modern browsers are restrictive about mixed content, and getting our secure frontend to communicate reliably with a local persistence layer (http://localhost) was a battle. We had to architect a proxy solution that satisfied browser security standards while allowing local development, ensuring seamless communication between our client, server, and the external AI APIs. Additionally, balancing state preservation across different learning modes without data loss required complex state management that we had to build from scratch.

Accomplishments that we're proud of

  • Total Proficiency Awareness: We successfully built a model that analyses speaking and writing to give users a clear, numerical understanding of their current standing, removing the guesswork from self-study.
    • Actionable AI Coaching: Our feedback system goes beyond simple grading; it provides the specific linguistic guidance needed to bridge the gap between a user's current level and the next milestone.
    • The Guided Roadmap: We’ve created a true step-by-step path that transforms passive consumption into a structured multimodal journey, ensuring every minute spent practising leads to the next level of mastery.
    • The Motivation Loop: By integrating a daily streak system with real-time visual progress, we’ve made the habit of consistent daily practice both rewarding and addictive.

What we learned

  • AI as a Supportive Friend: We discovered how AI can act as a daily companion, providing the instant validation and encouragement necessary to stay consistent with practice.
  • Decoding Proficiency: We gained deep insights into global standards and how to translate complex speaking and writing patterns into clear, actionable levels.
  • The Power of Awareness: We learned that knowing your exact current standing is the most powerful motivator for continuous improvement.
  • Direction Over Effort: We realised that a lack of direction is a bigger obstacle than a lack of effort; having a guided path transforms hard work into meaningful progress.

What's next for FluencyPath

The journey doesn't stop here. The roadmap for FluencyPath includes:

  1. Educational Partnerships: Collaborating with schools and universities to integrate FluencyPath into curriculums, helping students build real-world communication confidence alongside their academic studies.
  2. Real-time AI Conversation: Moving from recorded audio to live, low-latency voice chats with an AI tutor for immersive practice.
  3. Gamified Social Streaks: Implementing study groups where friends can compete on daily goals and celebrate milestones together.
  4. Video Analysis: Utilizing advanced multimodal capabilities to analyze body language and non-verbal cues during mock interviews.

Built With

  • express.js
  • filereader-api
  • firebase-admin-sdk
  • google-firestore
  • google-gemini-api
  • localstorage
  • mediarecorder-api
  • node.js
  • node.js-crypto
  • picsum-photos
  • pollinations.ai
  • react
  • react-markdown
  • remark-gfm
  • svg
  • tailwind-css
  • typescript
  • vite
Share this project:

Updates