Voice Assistant

This project is a voice assistant that uses FastAPI for the backend and Streamlit for the frontend. It can listen to your voice commands, transcribe them, generate responses using OpenAI's GPT-4, and provide functionalities like checking the weather and sending emails.

Features

Weather Information: Get real-time weather updates for any city using the OpenWeatherMap API.
Email Sending: Send emails directly through the assistant using the Resend API.
Voice Interaction: Speak your commands, and the assistant will listen, transcribe, and respond.

Setup

1. Clone the Repository

git clone https://github.com/yourusername/voice-assistant.git
cd voice-assistant

2. Set Up Environment Variables

Ensure you have the following environment variables set in your shell: On Mac/Linux:

export OPENWEATHERMAP_API_KEY='your_openweathermap_api_key'
export RESEND_API_KEY='your_resend_api_key'
export OPENAI_API_KEY='your_openai_api_key'

On Windows:

set OPENWEATHERMAP_API_KEY=your_openweathermap_api_key
set RESEND_API_KEY=your_resend_api_key
set OPENAI_API_KEY=your_openai_api_key

3. Install Dependencies

Ensure you have Python installed, then install the required Python packages:

pip install -r requirements.txt

4. Run the Backend

Start the FastAPI backend server:

fastapi dev app.py

5. Run the Frontend

In another terminal window, run the Streamlit frontend:

streamlit run frontend.py

API Endpoints /get_weather/{city}

Fetch the weather for a specified city.

Method: GET
Parameters:
    city (string): Name of the city.
Response:
    status: "success" or "failed"
    temp: Temperature in Celsius
    description: Weather description

/send_email

Send an email to a specified address.

Method: POST
Request Body:
    email_account (string): Receiver's email address.
    subject (string): Email subject.
    message (string): Email content.
Response:
    status: "success" or "failure"

/generate_response

Generate a response from the GPT-4 model using voice transcriptions.

Method: POST
Request Body:
    model (string): The OpenAI model to use (default: "gpt-4o").
    messages (list): List of messages (roles: "user" or "assistant").
Response:
    Returns the updated message history, including the assistant's response.

How It Works

Frontend (Streamlit): Records your voice, detects speech using WebRTC VAD, and transcribes it using OpenAI's Whisper model. The transcription is then sent to the backend.
Backend (FastAPI): Processes the transcription, calls the OpenAI API to generate a response, and optionally performs actions like fetching weather data or sending an email.
Response Handling: The assistant's response is sent back to the frontend, converted to speech, and played back to you.

Environment Variables

OPENWEATHERMAP_API_KEY: API key for OpenWeatherMap.
RESEND_API_KEY: API key for Resend.
OPENAI_API_KEY: API key for OpenAI.

Commands to Set Environment Variables

Mac/Linux

export OPENWEATHERMAP_API_KEY='your_openweathermap_api_key'
export RESEND_API_KEY='your_resend_api_key'
export OPENAI_API_KEY='your_openai_api_key'

Windows

set OPENWEATHERMAP_API_KEY=your_openweathermap_api_key
set RESEND_API_KEY=your_resend_api_key
set OPENAI_API_KEY=your_openai_api_key

Contributing

Contributions are welcome! Feel free to open an issue or submit a pull request. License

This project is licensed under the MIT License.

RediatBrook/voice-assistant