SpeakEasy

SpeakEasy is a language learning platform that focuses on immersing users in real-world scenarios. Each scenario is conducted through actual conversations, where the user speaks to our application about the specific scenario. The scenarios are designed to provide contextual learning, where the flow of conversation is dynamic, driven by interaction between the user and a generative AI. This allows for personalised and adaptive language practice, offering a more natural and engaging learning experience.

Application: https://speakeasy-speakfreely.netlify.app/

Group Members

Daniel Kok

• Matric Number: A0234683Y
• Contributions:
  • Designed frontend on Figma
  • Developed and designed landing page
  • Developed frontend application pages

Justin Tan

• Matric Number: A0252153N
• Contributions:
  • Prompt Engineering with Langchain
  • Integrate LLM for AI Responses, Context Refinement, and User Feedback
  • Creation of Fixed Scenarios
  • LLM and API Testing

Lam Jiu Fong

• Matric Number: A0255851Y
• Contributions:
  • Set up MongoDB
  • Developed Frontend pages
  • Wrote APIs in Frontend to link Backend
  • Set up Speech Recognition
  • Set up Text-to-Speech

Bryan Ho

• Matric Number: A0234509E
• Contributions:
  • Wrote backend APIs
  • Hosted Frontend on Netlify and Backend on Vercel
  • UI enhancements
  • Set up Google Authentication
  • Set up and implement Google Cloud Storage

Set Up Instructions

Frontend

• At the root of the repository, run npm i to install the packages
• Run npm run dev to start the application

Backend

• Change directory to backend folder and run python3 -m venv env to create the virtual environment

• Then run source env/bin/activate to enter the virtual environment

• The run pip install -r requirements.txt to download the dependencies

• Change directory to project and run python3 manage.py runserver to start the backend

  Note:

  Environment variables required for the Frontend

  • VITE_BACKEND_URL="http://127.0.0.1:8000"
  • VITE_GOOGLE_APIKEY
  • VITE_GOOGLE_AUTHDOMAIN
  • VITE_GOOGLE_PROJECTID
  • VITE_GOOGLE_STORAGEBUCKET
  • VITE_GOOGLE_MESSAGINGSSENDERID
  • VITE_GOOGLE_APPID
  • VITE_GOOGLE_MEASUREMENTID

  Environment variables required for the Backend

  • OPENAI_API_KEY
  • GOOGLE_APPLICATION_CREDENTIALS_JSON
  • MONGODB_KEY

Resources

Frontend

• TypeScript
• React
• ShadCN
• TailwindCSS
• Firebase Authentication
• OGP

Backend

• Python
• Django
• MongoDB
• Google Cloud Storage for blob store
• Langchain
• OpenAI LLM model