Welcome to Tandembot!

A voice chatbot utilizing speech recognition, speech synthesis, and artificial intelligence through the Google Gemini API to interact with users.

• Features
• Prerequisites
• Installation
• Usage
• Example Usage
• Future Improvements
• Contributing
• License

Features

• Recording of user input audio.
• Speech recognition to convert user input into text.
• Interaction with an artificial intelligence model to generate responses.
• Speech synthesis to transform chatbot responses into audio.

Prerequisites

• Python 3.x
• Python libraries listed in requirements.txt
• Valid credentials for the Gemini API

Installation

1. Clone the repository:

git clone https://github.com/anacletu/virtual_tandem

2. Install dependencies:

pip install -r requirements.txt

3. Create a .env file in the project root and add your API key, endpoint, and audio preferences:

API_KEY=your_api_key
API_URL=https://generativelanguage.googleapis.com/v1beta/models/gemini-pro:generateContent
FS=44100
DURATION=15

Usage

Run the main script:

python main.py

Follow the instructions in the terminal to interact with the chatbot. The script will make a request to the API, convert the response into speech, and play the audio.

Example Usage

See a quick video demonstrating simple conversations in Portuguese, English, and Spanish.

Tandembot.mp4

Future Improvements

• Addition of support for more languages.
• Implementation of a graphical interface to facilitate interaction.
• Improvement of speech recognition and speech synthesis robustness.
• More configuration possibilities, such as language level and response complexity.

Contributing

Contributions are welcome! For suggestions, bug fixes, and other changes, feel free to open an issue or submit a pull request.

License

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