One Api

🚀 One API Project Documentation

One API is an open - source platform based on large language models (LLMs), supporting multiple mainstream AI models. It connects multiple different LLM providers through a unified interface and offers a rich set of features to manage and optimize the calls to these models.

🚀 Quick Start

Environment Requirements

• Operating System: Linux (recommended) or macOS
• Programming Language: Python 3.8+
• Dependencies:
  • python - dotenv: Used to load environment variables.
  • websockets: A library supporting the WebSocket protocol.
  • fastapi: An ASGI - based Web framework.
  • uvicorn: An ASGI server.

Installation and Deployment

Clone the Repository

git clone https://github.com/yourusername/one-api.git
cd one-api

Install Dependencies

pip install python-dotenv websockets fastapi uvicorn

Start the Service

uvicorn main:app --reload

Configuration File

The project uses the config.json file for configuration. Here is an example:

{
  "debug_mode": true,
  "web_port": 8000,
  "allowed_origins": ["http://localhost", "https://yourdomain.com"],
  "api_key": "your_api_key_here"
}

✨ Features

1. Multi - model Support: Integrates multiple mainstream LLM providers.
2. Channel Management: Allows configuration and management of different API providers.
3. Quota Calculation: Calculates fees based on the consumption of prompt tokens and completion tokens.

💻 Usage Examples

Online Demo

Visit the following link to experience it: Online Demo Site

Screenshots

📚 Documentation

Frequently Asked Questions

1. What is the quota and how is it calculated?

   • Quota = Group multiplier × Model multiplier × (Number of prompt tokens + Number of completion tokens × Completion multiplier)
   • The multipliers for prompt tokens and completion tokens vary by model.

2. Questions about API calls

   • Ensure that the api_key in the configuration file is correct.
   • Check the network connection to ensure access to external services.

3. How to handle errors?

   • Check the detailed log information in the console output.
   • Confirm that the environment variables and configuration file are set correctly.

Command - line Options

Start Command

uvicorn main:app --reload

Other Parameters

• --port: Specify the service port, default is 8000.
• --debug: Enable debug mode to display detailed logs.

Related Projects

• GitHub Repository
• Official Documentation

📄 License

1. MIT License: This project follows the MIT open - source license, allowing free use, modification, and distribution.
2. Attribution Requirement: In second - development projects, please indicate the source and keep the relevant links.
3. Issue Feedback: Welcome to submit issues or pull requests on GitHub.

If you have any questions or need further assistance, please refer to the Official Documentation.