oproxy 1.0.0

High-performance, transparent HTTP/HTTPS proxy that supports both TCP and UDP protocols while being fully configurable and privacy-focused

OProxy

High-performance, transparent proxy that supports both TCP and UDP protocols.

A high-performance, transparent proxy that supports both TCP and UDP protocols.

NOTE: This proxy has been designed with local API proxy in mind. Specifically, I used it to forward Ollama API requests to the remote Ollama server for applications that try to connect to the local Ollama server on localhost.

Features

• Transparent TCP proxying
• HTTP/HTTPS proxying without decrypting the traffic
• Headers and other metadata fully preserved
• Optional UDP support
• Detailed logging capabilities
• Configurable through environment variables
• Support for both file and stdout logging
• Data content logging (optional)
• Performance optimizations with configurable buffer sizes
• Real-time metrics monitoring
• Automatic log rotation
• Thread-safe metrics collection
• Throughput and connection statistics
• Supports standalone usage as well as being used as a Python module

IMPORTANT: The proxy will automatically log stuff to /var/log/oproxy.log if no log file is specified. Please make sure the user running the proxy has write permissions to this file or specify a different log file using the --log-file argument.

Requirements

• Python 3.7+
• python-dotenv
• socket
• threading

To build the package

• hatchling
• setuptools
• wheel
• build

Performance Features

Metrics Monitoring

The proxy now includes built-in metrics collection and monitoring:

• Total connections tracking
• Active connections monitoring
• Bytes transferred counting
• Real-time throughput calculation
• Periodic metrics reporting (every 60 seconds)

Performance Optimizations

• Optimized buffer sizes (65KB)
• Non-blocking I/O using select
• Socket buffer optimization
• Thread-safe operations

Log Management

• Automatic log rotation (10MB per file)
• Up to 5 backup log files
• UTF-8 encoding support
• Compressed backup files

Installation

From source

1. Clone the repository:

git clone https://github.com/tcsenpai/oproxy.git

cd oproxy

2. Install dependencies:

pip install -r requirements.txt

3. Copy the example environment file:

cp .env.example .env

4. Edit the .env file with your configuration:

# Example: your Ollama server is running on 192.168.1.100:11434
PROXY_PORT=11434
TARGET_HOST=192.168.1.100
TARGET_PORT=11434

From PyPI

pip install oproxy

NOTE: Once installed through PyPI, you can use the oproxy command to start the proxy instead of python src/main.py. This is valid also for the following usage examples.

Usage

Basic TCP proxy:

python src/main.py

Customize the proxy port and/or target host and port:

python src/main.py --proxy-port 11435 --target-host 192.168.1.101 --target-port 11435

Enable logging to file:

python src/main.py --log-file=proxy.log

Enable data logging with debug level:

python src/main.py --log-file proxy.log --log-data --log-level DEBUG

Enable full data logging:

NOTE: This will log the entire payload of the request and response.

python src/main.py --log-file proxy.log --log-data --full-debug

Enable UDP support:

python src/main.py --enable-udp

View Metrics

Metrics are automatically logged to your configured log file or stdout. They include:

Performance Metrics: {
    'total_connections': 150,
    'active_connections': 3,
    'bytes_transferred': 1048576,
    'uptime_seconds': 3600,
    'bytes_per_second': 291.27
}

Command Line Arguments

• --proxy-port: Port the proxy will listen on (default: 11434)
• --target-host: Target host to forward traffic to (default: 192.168.1.100)
• --target-port: Target port to forward traffic to (default: 11434)
• --log-file: Path to the log file
• --log-data: Enable logging of data content
• --log-level: Set logging level (DEBUG, INFO, WARNING, ERROR, CRITICAL)
• --enable-udp: Enable UDP proxy alongside TCP
• --full-debug: Enable full data logging (entire payload)

Notes

• TCP proxy runs on the port specified in .env
• UDP proxy (if enabled) runs on PROXY_PORT + 1
• Data logging should be used carefully as it may contain sensitive information
• UDP support is experimental and runs as a daemon thread
• HTTPS proxying is handled without decrypting the traffic

License

Apache-2.0 License. See the LICENSE file for details.