openai-realtime-fastapi

This project facilitates real-time communication between FastAPI WebSocket connections and OpenAI's WebSocket connections, including compatibility with Azure OpenAI. It acts as a relay server that can be extended for security and customization purposes.

Table of Contents

• Overview
• Setup and Installation
• Configuration
• Usage
• Development
• Benefits and Use Cases
• Acknowledgments

Overview

This application uses FastAPI to provide a WebSocket interface that connects with OpenAI's WebSocket API. The project is structured to handle real-time data exchange, making it suitable for scenarios where latency and live updates are crucial. It also supports integration with Azure OpenAI.

Setup and Installation

Prerequisites

• Python 3.8+
• An OpenAI API Key or Azure OpenAI API Key

Installation

1. Clone the repository:

   git clone https://github.com/Geo-Joy/openai-realtime-fastapi
   cd openai-realtime-fastapi

2. Set up a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install dependencies:

   pip install -r requirements.txt

4. Create a .env file in the root directory with the following content. Use the appropriate URL depending on whether you are using OpenAI or Azure OpenAI:

   OPENAI_API_KEY="your-openai-api-key"
   OPENAI_REALTIME_URL="wss://api.openai.com/v1/realtime?model=gpt-4o-realtime-preview-2024-10-01"
   USE_AZURE_OPENAI=False

   For Azure OpenAI, replace the URL as follows and set USE_AZURE_OPENAI to True:

   OPENAI_API_KEY="your-azure-openai-api-key"
   OPENAI_REALTIME_URL="wss://<sub-domain>.openai.azure.com/openai/realtime?api-version=2024-10-01-preview&deployment=gpt-4o-realtime-preview"
   USE_AZURE_OPENAI=True

Configuration

Configuration settings are managed using environment variables loaded from a .env file. The main configuration file is config.py, which retrieves values like API_KEY and VENDOR_WS_URL from the environment.

Usage

To start the application, use Uvicorn to run the FastAPI server:

uvicorn main:app --reload

The application should now be running on http://localhost:8000.

Development

Key Files

• main.py: Initializes the FastAPI app and includes routers for various endpoints.
• health_check.py: Provides a simple health check route.
• realtime.py: Manages the WebSocket connection and data relay logic.

Benefits and Use Cases

The openai-realtime-fastapi project can act as a relay server with various advantages:

1. Security and Privacy:

   • API Credential Protection: By routing requests through a server, API keys are kept secure and hidden from client-side exposure.
   • Sensitive Data Handling: Execute sensitive operations server-side to prevent leakage or manipulation from client-side interactions.

2. Controlled Environment:

   • Centralized Logic: Define and manage business logic server-side, ensuring uniform application behavior and more straightforward updates.
   • Restricted Access: Permit only specific types of requests and responses between the client and third-party services, enhancing security and rule enforcement.

3. Authentication and Authorization:

   • Flexible Authentication Mechanisms: Implement various authentication strategies (e.g., OAuth, JWT, API Key) based on application requirements to ensure secure client-server interactions.
   • Single Sign-On (SSO): Enable SSO to streamline user access, reducing the need for multiple login credentials, while keeping vendor connections and API interactions hidden from the end-user.

4. Reduced Client Complexity:

   • Clients can interface with a simplified API rather than directly interact with complex third-party services.
   • Streamline client-side applications since logic and API integrations are managed server-side.

5. Flexibility for Tests and DevOps:

   • Mock or alter API responses for different environments without impacting production clients.
   • Seamlessly switch between different service providers or deployments.

Acknowledgments

• FastAPI
• OpenAI API
• Azure OpenAI
• Contributor: Geo Joy (https://www.linkedin.com/in/-itsg/)