Skip to content

README.en.md

Latest commit

 

History

History
209 lines (156 loc) · 6.3 KB

README.en.md

File metadata and controls

209 lines (156 loc) · 6.3 KB
Raw

Cursor-To-OpenAI-Nexus

中文 | English

Forward Cursor API requests to OpenAI, with support for multiple API Keys rotation.

Features

  • 🔑 Multiple Keys Rotation: Configure multiple API Keys rotation to improve availability
  • 🚀 Easy Configuration: One-click configuration script for quick setup
  • 📊 Status Monitoring: Monitor API Key usage status
  • 🔧 Easy Maintenance: Convenient maintenance scripts to simplify daily operations

🚀 Basic Installation

Clone Repository

git clone https://github.com/liuw1535/cursor-to-openai-nexus.git

Enter Project Directory

cd cursor-to-openai-nexus

Install Dependencies

npm install

⚙️ Configure Project

npm run setup
  • Just fill in the custom key and whether to enable TLS proxy server
  • Other options can be skipped by pressing Enter or filled in randomly
  • 🛡️ If you frequently encounter account blocking issues, it's recommended to enable TLS server
  • If you're not satisfied with the configuration, you can re-run this command to modify it

🏃 Start Service

npm start

🔍 Usage

  1. Access the management interface: http://127.0.0.1:3010
  2. Use the blue button at the bottom of the page to get cookies
  3. Configure in the Tavern page:
    • API address: http://127.0.0.1:3010/v1
    • Key: sk-text (if "text" was entered during configuration)

📧 Account Registration Recommendations

  • Recommended to use domain email (subdomain email is better)
  • Search for "cloudfare domain email" for configuration tutorials
  • ⚠️ Register no more than 2 accounts at a time to avoid being blocked

🛠️ Common Commands

npm start           # Start project
npm run setup       # Modify configuration

Environment Configuration

Configure the following key parameters in the .env file:

  • API_KEYS: Mapping relationship between API Key and Cookie (JSON format)
  • USE_TLS_PROXY: (true) Enable TLS server, which can avoid request blocking issues
  • PROXY_PLATFORM: The platform corresponding to the TLS server when enabled, default is auto detection

The system will automatically merge API Keys from .env and data/api_keys.json at startup to ensure data consistency.

Deployment Method

Using Docker Compose

# Create configuration files
cp .env.example .env
mkdir -p data
cp data/admin.example.json data/admin.json

# Create admin account
node scripts/create-admin.js

# Start service
docker compose up -d --build

# View logs
docker compose logs -f

# Stop service
docker compose down

API Usage Example

Python Example

from openai import OpenAI

# Use custom API Key
client = OpenAI(api_key="your_custom_api_key",
                base_url="http://localhost:3010/v1")

# Or use Cookie directly
# client = OpenAI(api_key="user_...",
#                base_url="http://localhost:3010/v1")

response = client.chat.completions.create(
    model="claude-3-7-sonnet",
    messages=[
        {"role": "user", "content": "Hello."},
    ],
    stream=False
)

print(response.choices)

Notes

  • Please keep your WorkosCursorSessionToken secure
  • This project is for learning and research purposes only, please comply with Cursor's terms of use

Acknowledgements

Logging System

The project integrates a unified logging system, which can be configured through the following methods:

Log Level Configuration

  1. Set environment variables in the .env file 
    LOG_LEVEL=INFO
LOG_FORMAT=colored
LOG_TO_FILE=true
LOG_MAX_SIZE=10
LOG_MAX_FILES=10
  2. Specify environment variables in the startup command, for example: LOG_LEVEL=DEBUG npm start

Supported log levels include:

  • ERROR: Only display error messages
  • WARN: Display warning and error messages
  • INFO: Display general information, warnings, and error messages (default)
  • DEBUG: Display debug information, general information, warnings, and error messages
  • TRACE: Display all log information

Log Format

The log format is: [LEVEL] timestamp log content, with different levels displayed in different colors for easy differentiation:

  • ERROR: Red
  • WARN: Yellow
  • INFO: Green
  • DEBUG: Blue
  • TRACE: Cyan
  • HTTP: Cyan (dedicated to HTTP request logs)

HTTP Request Logs

The project uses the Morgan middleware to record HTTP requests, integrated into the unified logging system:

  1. Set HTTP log format in the .env file:

    # Options: tiny, combined, common, dev, short
MORGAN_FORMAT=tiny

  2. HTTP logs will be displayed with the [HTTP] prefix, highlighted in cyan for easy identification

  3. Morgan format options explanation:

    • tiny: The most concise format, including only method, URL, status code, response time
    • combined: Standard Apache combined log format, including IP, time, request, status code, response size, referrer, user-agent
    • common: Standard Apache common log format, similar to combined but without referrer and user-agent
    • dev: Developer-friendly colored format, including method, URL, status code (with color), response time
    • short: Shorter format, including method, URL, status code, response time, response size

File Logs

The project supports outputting logs to both console and files, which can be enabled with the following configuration:

  1. Set in the .env file: LOG_TO_FILE=true
  2. Optional configuration:
    • LOG_MAX_SIZE: Maximum size of log file, in MB, default 10MB
    • LOG_MAX_FILES: Number of historical log files to keep, default 10

Log files are stored in the logs folder in the project root directory:

  • Current log file: app.log
  • Historical log file: app-2023-05-05T12-45-30-000Z.log

File logs will automatically rotate, creating a new log file when the log file size exceeds the set value and keeping the most recent N files.

Usage in Code

Different log levels can be used as needed in the code:

const logger = require('./utils/logger');

logger.error('This is an error message');
logger.warn('This is a warning message');
logger.info('This is a general information message');
logger.debug('This is a debug message');
logger.trace('This is a trace message');
logger.http('This is an HTTP request log');