Virtual Assistant with Gemini API

Objective

This project demonstrates how to use the Gemini API to build your own virtual assistant. The assistant responds to a wake word, processes user input, and provides intelligent answers. This project is designed as a hands-on tutorial for developers looking to learn how to integrate the Gemini API into their own projects, with a focus on voice interaction.

Features

Wake Word Activation: The assistant listens for the wake word "Hey ADA" before processing any input. Voice Recording: Once activated, the assistant records the user's voice, sends the audio input to the Gemini API, and receives a text response. Custom Responses: Based on user prompts, the assistant provides intelligent responses in both text and voice formats.

Requirements

Python 3.x Gemini API Access Libraries: pyaudio, speech_recognition

How It Works

Wake Word Detection: The virtual assistant continuously listens for the phrase "Hey ADA". Once heard, the assistant starts recording the user's voice. Processing: The recorded voice is sent to the Gemini API, where the user's query is processed. Response: The assistant receives a text response from the API, which is then converted to a voice response for the user.

Installation

Clone the repository:

git clone https://github.com/your-username/your-repo-name.git

Install the required dependencies:

pip install -r requirements.txt

Set up your Gemini API key in the environment variables:

GEMINI_API_KEY='your-api-key'

Run the application:

python app1.py

Files

• app1.py: Main application file that integrates wake word detection, voice recording, and communication with the Gemini API.
• grabar_audio2.py: Handles the audio recording and processing.
• on_start.py: Initializes the application and sets up necessary configurations.

Contribution

Feel free to contribute by opening issues or submitting pull requests.