This project provides an AI-powered IELTS Speaking simulator designed for individuals preparing for the IELTS exam. It offers an authentic practice experience by mimicking the exam's flow, randomly selecting up-to-date questions, and providing post-conversation feedback and text improvement suggestions.

How It Works

The simulator leverages ChatGPT's GPTs and voice capabilities to replicate the IELTS speaking test. It utilizes a custom-built external API for stable random question selection, overcoming limitations of ChatGPT's internal Code Interpreter for this task. The system simulates the exam's three parts, offers automated scoring (with caveats on pronunciation accuracy), and prompts users for answer optimization and vocabulary summarization.

Quick Start & Requirements

• Access requires the official ChatGPT application (iOS/Android or macOS with M1+ chip).
• Users need a ChatGPT account; free users have limitations, while ChatGPT Plus is recommended for full practice.
• Installation involves accessing the GPT via a provided link or searching within the ChatGPT app and saving it to the sidebar.
• For optimal experience, set the ChatGPT app's main language to English.

Highlighted Details

• Updated with the latest 2024 May-August IELTS question bank.
• Enhanced stability in random question drawing via an external API.
• Automatic prompts for answer improvement post-simulation.
• Supports English and Chinese (Simplified/Traditional) languages.

Maintenance & Community

The project is actively maintained by hubeiqiao, with updates reflecting new question banks and feature optimizations. The creator shares the GPT's instructions and question banks, encouraging community contributions and modifications.

Licensing & Compatibility

The project's licensing is not explicitly stated in the README. Compatibility is tied to OpenAI's ChatGPT platform and its availability on various devices.

Limitations & Caveats

Pronunciation scoring is noted as inaccurate due to current voice-to-text limitations; future GPT-4o features may improve this. Part 1 experience can be inconsistent, sometimes exceeding the question limit or repeating questions. The system may occasionally miss questions or provide incomplete answers, requiring user correction. Voice functionality can be unstable depending on network conditions.