OfferGoose Clone

This project is an Interview Assistant App that converts the interviewer's spoken questions to text using a speech-to-text package and fetches real-time responses from a Gemini-tuned model. The application is built using the Flutter framework and leverages the Bloc Cubit for state management. It follows the Repository, Observer, and Dependency Injection design patterns for clean and scalable architecture.

Features

• Speech-to-Text Conversion: Converts the interviewer's spoken questions into text.
• Real-Time AI Responses: Fetches responses using the Gemini-tuned model API in real-time.
• State Management: Implements Bloc Cubits to manage states efficiently across the app.
• Repository Pattern: Decouples business logic from the UI by using repositories.
• Dependency Injection: Manages dependencies using DI, ensuring flexibility and testability.
• Observer Pattern: Monitors and handles application state transitions.

Screens

1. Home Screen:

   • Entry point of the app, where the user can initiate the interview process.

2. Formal Interview Screen:

   • Displays the speech recognition interface, capturing the interviewer's question and converting it into text.

3. Response Screen:

   • Displays the real-time AI-generated response based on the recognized question, streaming the response from the Gemini-tuned model.

Architecture Overview

• Cubit State Management:

  • The app uses Cubit for managing various states such as speech recognition, question response streaming, and timers.

• Repository Pattern:

  • A FormalInterviewResponseRepository is used to interact with the Gemini model API and fetch responses. This ensures a clean separation between the data layer and business logic.

• Dependency Injection (DI):

  • Dependencies such as the FormalInterviewResponseRepository and various Cubits are injected throughout the app to maintain testability and flexibility.

• Observer Pattern:

  • The app observes state transitions within the Cubits, handling errors, loading, and response events effectively.

Dependencies

• Flutter Bloc: For state management using Cubit.
• Speech to Text Package: To capture and convert the interviewer's speech into text.
• Gemini API (via OpenAI API): To generate responses in real-time.
• Dart dotenv: For environment variable management.
• HTTP: For making network requests to the Gemini API.

Dependencies

The application utilizes the following packages:

• cupertino_icons: ^1.0.8
• flutter_screenutil: ^5.9.3
• speech_to_text: ^7.0.0
• flutter_spinkit: ^5.2.1
• flutter_bloc: ^8.1.6
• flutter_dotenv: ^5.1.0
• google_generative_ai: ^0.4.6

Installation

1. Clone the Repository:

   git clone https://github.com/yourusername/interview-assistant-app.git

2. Navigate to the Project Directory::

cd interview-assistant-app

3. Install Dependencies:

flutter pub get

4. Set up the Environment Variables::

OPEN_AI_API_KEY=your_gemini_api_key_here

5. Run the App:

Run the App

How It Works

Speech-to-Text Recognition:

• The user speaks the question into the app on the Formal Interview Screen.
• The SpeechRecognitionCubit listens for the speech input and converts it into text using the speech-to-text package.

Fetching AI Response:

• Once the question is recognized, the app sends the text to the Gemini-tuned model via the FormalInterviewResponseCubit.
• The response is streamed in real-time, updating the Response Screen with each portion of the answer as it arrives.

State Management:

• Each screen is managed by a corresponding Cubit, which handles state transitions (e.g., loading, streaming, errors).
• The app follows the Bloc Observer pattern to manage these states and log errors.

Contribution Guidelines

1. Fork the repository.
2. Create a new branch for your feature or bug fix.
3. Submit a pull request, detailing the changes you've made.

License

This project is licensed under the MIT License. See the LICENSE file for more details.

Conclusion

Thank you for exploring the Interview Assistant App. We hope this application serves as a valuable tool for interview preparation and enhances your experience by leveraging advanced speech recognition and AI response capabilities. For any queries, feedback, or contributions, feel free to reach out.