QuickLesson

A web application that makes language speaking practice simple through short, real-time conversations with native speakers.

🌐 Live Demo: https://django-5min-languageapp.onrender.com/ 📄 Status: Active Development

---

Overview

QuickLesson is a language exchange platform designed for learners who want to improve their speaking skills without committing to long lessons.

Instead of one-hour sessions, users join 5-minute conversations with native speakers whenever they have time.

---

Features

• 🌍 Multiple languages (Japanese, English, Spanish, French)
• 🎥 Real-time video calls
• 👩‍🏫 Native speaker matching
• ⭐ Tutor ratings
• 📖 Session history
• 👤 User profiles
• 📱 Responsive design

---

Tech Stack

Backend

• Django
• Django REST Framework
• PostgreSQL

Frontend

• HTML
• CSS
• JavaScript

Deployment

• Render

---

Screenshots

Home	point

---

Project Structure

quicklesson/
├── accounts/
├── lessons/
├── tutors/
├── chat/
├── static/
├── templates/
└── manage.py

---

Installation

git clone ...
cd quicklesson
pip install -r requirements.txt
python manage.py migrate
python manage.py runserver

---

Roadmap

• Authentication
• Video calls
• Tutor dashboard
• AI feedback
• Mobile app
• Push notifications

---

About the Project

QuickLesson is a personal project that I continue to develop while collecting feedback from real users.

The goal is to make language speaking practice accessible, flexible, and enjoyable through short conversations.

Next-conversation questions

After a tutor saves a note, the server generates five questions using OpenAI's Responses API (gpt-4.1-mini by default). Only the same student's latest five notes in the lesson language are included. The current CEFR level, or the latest specified level in that history, controls difficulty. No account names or emails are added to the request. Note text itself is sent to OpenAI; store is false.

Questions are saved on the note and shown below the tutor guide in the next lesson. Opening the room never generates questions. Repeated submissions for the same match reuse the first saved note and do not generate again. API failures, missing keys, or missing levels leave the note saved without suggestions; no automatic retry is performed. A save may wait up to the HTTP timeout (8 seconds).

Local development reads OPENAI_API_KEY from the ignored .env.local file if it is not already in the environment. Production requires OPENAI_API_KEY in the hosting service's secret environment settings. Never commit the env file or key. OPENAI_QUESTION_MODEL may override the model. Run python manage.py migrate after deployment to add the question fields (the existing build script does this). Existing notes are not automatically sent to the API or backfilled.