dialekt

Dialekt is an innovative AI-powered language tutor designed to enhance language learning through immersive, adaptive pedagogy aligned with the Common European Framework of Reference for Languages (CEFR). This project employs an intelligent conversational interface that facilitates learning in real-time, ensuring an engaging experience for users of various proficiency levels, from A1 to C2.

Key Features

• Adaptive Conversations: The system adapts vocabulary and grammar according to the user’s CEFR level, ensuring an appropriate challenge.
• Extensive Language Support: Supports over 70 languages, including major languages such as English, Spanish, French, German, Mandarin, Japanese, Portuguese, Arabic, Russian, and Hindi.
• Flexible LLM Providers: Users can choose from multiple language model providers such as Anthropic, OpenAI, or OpenRouter for personalized interactions.
• Real-Time Feedback: Ensures contextual learning through immediate corrections, translations, and International Phonetic Alphabet (IPA) phonetics.
• Text-to-Speech Technology: Provides synthesized native pronunciation for all supported languages, improving listening skills.
• Voice Input Functionality: Enables hands-free language practice via a browser-based speech-to-text feature, allowing natural speech without typing.
• Session Management: Users can save and resume their language learning configurations seamlessly.

How It Works

Users can type in either their native or target language, with the AI tutor always responding in the target language to maintain immersion. If a user writes in their native language, the message is translated with phonetic guidance. When practicing the target language, immediate corrections and constructive feedback are provided. Each interaction includes translations and phonetics to aid comprehension.

System Architecture

The architecture includes a user input layer where users interact in their chosen languages, an adaptive business logic layer that handles CEFR prompt creation and maintains register rules, and an HTTP client layer that connects to various language models for processing user input. The structured response format ensures clarity and aids in effective learning by presenting responses along with translations and follow-up questions.

Getting Started

To initiate Dialekt, users can set up the PostgreSQL database and install dependencies, followed by configuring the environment to connect to the selected LLM provider. After everything is set up, the application is served locally for access.

Voice Input Workflow

Dialekt's voice input feature allows users to record their speech for natural language practice. The workflow is intuitive: start recording, switch languages seamlessly, and review the transcribed text before sending it. This hands-on practice enhances fluency and confidence in language use.

In summary, Dialekt serves as a comprehensive language learning tool that combines advanced AI technologies with user-centered design to facilitate effective communication skills across diverse languages.