Chatterpals
Image credit: Unsplash“Foreign Language Learning Friends Who Chat with AI”
Overview
- Service Overview
- Problem Definition and Solutions
- Core Ideas and Features
- Expected Effects
- Demo Page Structure
- Utilized Datasets
- Expansion Directions
- Conclusion
🎯 Service Overview
ChatterPals is a foreign language learning service that goes beyond simple grammar study, helping users develop real conversational skills through real-time dialogue with AI.
When a user is viewing actual content such as news, blogs, social media, or YouTube, the on-screen information is read by an AI tutor character, which then asks questions or provides feedback.
In this way, learners can naturally converse and learn in a foreign language anytime, anywhere.
🧩 Problem Definition and Solutions
Existing Problems | ChatterPals’ Solutions
- AI learning focuses mainly on grammar and vocabulary, weak in real conversation → Provides real-time voice-based interactive tutoring
- Repetitive learning structure makes it boring and unsustainable → Offers engaging, trend-based conversational topics
- Lack of personalized feedback → Provides grammar, pronunciation, and intonation feedback PDFs using personal learning data analysis
- Lack of real-life speaking experience → Simulates real conversations through AI voice chat and virtual environments
ChatterPals maximizes the core abilities of AI tutors — real-time interaction, personalized feedback, and immersive learning — helping users develop true conversational fluency to speak naturally with real foreigners.
🧠 Core Ideas and Features
1️⃣ AI Tutor Character Popup
- A cute AI learning partner character appears on the user’s screen
- Reads the context of the webpage, article, or social media post the user is viewing
→ Asks questions like “What do you think about this article?”
- The user answers in a foreign language, and the AI immediately provides grammar and pronunciation corrections with alternative expressions
- Learning records are automatically saved and offered as a monthly personalized feedback report (PDF)
2️⃣ Learning Records and Statistics
- Categorizes learning content by topic and platform (e.g., society, travel, hobbies, reviews, daily life)
- Allows users to check learning frequency, duration, and progress
- Favorite feature for managing preferred content
3️⃣ Personalization and Tutor Customization
- Adjusts learning style based on the learner’s level and goals
- Users can customize the tutor’s personality, voice, and appearance
- Allows selection of feedback tone (encouraging, strict, etc.)
4️⃣ Level Test
- Evaluates users based on voice recognition and sentence construction
- Diagnoses current level and recommends a personalized learning path
🚀 Expected Effects
- Enhances real conversational ability
- Corrects pronunciation and intonation through AI voice interaction
- Provides real-time feedback for natural sentence formation
- Improves learning consistency
- Uses trending topics from SNS and news to maintain interest
- Motivates users with gamified elements such as character popups, points, and challenges
- Accessible anytime, anywhere
- Mobile and web synchronization enables spontaneous conversation practice
- Enables learning like “reading an article on the bus → discussing it in English with AI”
🔍 Demo Page Structure
- Home Screen : Entry point with four menus (Tutor, Settings, Level Test, Logs)
- AI Tutor : Core interactive learning function (discussion, explanation, rephrasing)
- Settings : Adjusts learning style, tutor character, voice, and feedback tone
- Level Test : Evaluates proficiency and recommends a personalized curriculum
- Learning Logs : Visualizes learning data and tracks personal progress
📊 Utilized Datasets
| Source | Dataset Name | Purpose |
|---|---|---|
| AIHub | Topic-Adaptive English Speaking Dataset for Koreans | Training real-time speaking evaluation models |
| AIHub | Korean-English Parallel Translation Corpus | Context-based translation and response modeling |
| AIHub | Multilingual Daily Conversation Translation Dataset | Handling multilingual voice conversations |
| Kaggle | GPT-4 Multilingual Dialogue Dataset | Learning natural conversation simulation |
| SelectStar | Multilingual Conversation Dataset | Training models to handle various languages and contexts |
🌐 Expansion Directions
Monthly Topic Challenges: AI-user discussions on topics like environment and sports Brand-Linked Learning Missions: Natural ad experiences within conversations VR/AR Conversation Simulations: Practice real-life situations like airports, restaurants, or interviews Community Learning System: Rankings, group discussions, and study sessions for enhanced social learning
💡 Conclusion
ChatterPals is a practical foreign language conversation learning platform that enables learners to chat and grow with AI.
Learn anytime, anywhere, and develop your real conversational skills with your adorable AI friends!
