Agador

Agador is a cutting-edge voice assistant application that combines speech recognition, AI-powered responses, text-to-speech synthesis, and face detection. The project leverages multiple APIs and libraries to create an interactive, voice-enabled experience.

Features

• Speech Recognition: Uses the Google Speech Recognition API to convert spoken language to text.
• Generative AI: Utilizes the Google Generative AI API for producing natural language responses.
• Text-to-Speech: Synthesizes audio responses using the ElevenLabs API.
• Face Detection: Monitors video streams through a webcam using OpenCV to detect faces.
• System Tray Integration: Provides a system tray icon for quick control and graceful shutdown.
• Audio Bypass Mode: Supports pre-recorded audio file input, bypassing the live microphone for testing and simulation.

Installation

1. Clone the Repository:

   git clone https://github.com/Heather-Herbert/agador.git
   cd agador

2. Install Dependencies:

   Ensure you have Python 3 installed. Install the project dependencies using pip and the provided requirements.txt:

   pip install -r requirements.txt

Configuration

.env File

Create a .env file in the project root directory to store your API keys and configuration settings. Below is a sample template:

# Google API Key for Generative AI (Get this from the Google Cloud Console)
GOOGLE_API_KEY=your_google_api_key_here

# ElevenLabs API Key (Obtain from your ElevenLabs account)
ELEVENLABS_API_KEY=your_elevenlabs_api_key_here

# Optional: ElevenLabs Voice ID (defaults to "pNInz6obpgDQGcFmaJgB")
ELEVENLABS_VOICE_ID=your_preferred_voice_id

# Optional: LLM model name (defaults to "gemini-pro")
LLM_MODEL_NAME=your_llm_model_name

# Optional: Minimum face size percent to trigger face detection (defaults to 10)
MIN_FACE_SIZE_PERCENT=10

# Optional: Microphone index to use (leave unset to use the default microphone)
MICROPHONE_INDEX=0

Obtaining API Keys:

• Google Generative AI:

  • Visit the Google Cloud Console.
  • Create a new project or select an existing one.
  • Enable the relevant Generative AI API.
  • Generate and copy your API key.

• ElevenLabs:

  • Sign up or log in at ElevenLabs.
  • Navigate to the API keys section in your account settings.
  • Copy your API key and add it to your .env file.

Audio Bypass Mode

Agador includes an audio bypass mode that allows the use of a pre-recorded audio file in place of the live microphone input. This mode is particularly useful for testing or simulating voice inputs.

How to Use Audio Bypass Mode

1. Prepare an Audio File:

   • Place your audio file (supported formats such as MP3 or WAV) in the data directory.
   • Rename the audio file to input_audio.mp3.

2. Run the Application:

   • When Agador starts, it checks the data folder for the input_audio.mp3 file.
   • If the file is found, Agador will process it in place of live audio captured from the microphone.
   • After processing, the file is renamed to prevent reprocessing in subsequent runs.

Running the Application

After installing dependencies and setting up your .env file, run the application with:

python agador.py

Agador will then launch with a system tray icon. The application will list available microphones (if using live mode) and continuously monitor for a face through your webcam to trigger the voice recognition flow.

Troubleshooting

• API Keys Missing or Invalid:
  Ensure your .env file is correctly configured with valid API keys.

• Microphone Issues:
  Verify that the MICROPHONE_INDEX is correct or leave it unset to display the option on your system.

• Audio File Handling:
  Confirm that your pre-recorded audio file is named input_audio.mp3 and located in the data directory if using audio bypass mode.

• Dependencies:
  Double-check that all required packages from requirements.txt are installed.

Contributing

Contributions are welcome! Whether you want to add new features, fix bugs, or improve documentation, please feel free to open an issue or submit a pull request.

License

This project is licensed under the MIT License.