AI-LLP

AI Language Learning Application
Developed for the DSAIT4065 Conversational Agents course at TU Delft.

👥 Authors

• Jordy del Castilho – jordydelcastilho@gmail.com
• Liwia Padowska – liwia.padowska@gmail.com
• Yizhen Zang – yizhenzang@tudelft.nl
• Zeryab Alam – zeryabalam272@icloud.com

🧠 Project Overview

AI-LLP is a conversational agent designed to assist with language learning. The system integrates multiple modules to simulate a realistic and helpful dialogue experience for learners.

📦 Modules

• ASR (Automatic Speech Recognition) – Converts spoken input to text.
• TTS (Text-to-Speech) – Converts text responses into speech.
• NLP (Natural Language Processing) – Understands and generates natural language responses.
• MEM (Memory) – Keeps track of user interactions and context.

The complete report detailing the design, methodology, and evaluation of this project is included in this repository.

🛠️ Installation & Setup

✅ Prerequisites

• Python 3.10+
  Download Python – tested primarily on Python 3.10.

• FFmpeg

  • Windows builds
  • macOS/Linux
    ⚠️ After installation, make sure ffmpeg is added to your system's environment variables.

• Ollama
  Install Ollama – used to run local language models.
  📌 Ensure you pull a model that fits your GPU memory constraints.

---

📥 Setup Instructions

1. Clone the repository: git clone git@github.com:zeri27/AI-LLP.git
2. Install dependencies: pip install -r requirements.txt

---

🚀 Usage

1. Ensure a stable internet connection (especially for the TTS module — use version v1 if offline).
2. Run the agent: python agent.py