​

How It Works

The proxy operates as a simple HTTP server that listens on a specified port. It is configured to intercept all traffic destined for api.openai.com and reroutes it—with the correct headers and configurations to oai.helicone.ai. This enables you to swiftly integrate Helicone and all of its features without having to alter a single line of code.

​

Potential Use Cases for the MITM Proxy

• Github Actions
  • Continuous Integration/Continuous Deployment (CI/CD)
  • Testing
  • Debugging
  • Cost savings (utilizing Bucket Cache)
• Local Development
  • Debugging
  • Testing
  • Cost savings (utilizing Bucket Cache)

​

Installation (Ubuntu/Debian)

We provide a script that will automatically install and configure the proxy for you.

​

Starting the Proxy

# Configure Cache
export HELICONE_CACHE_ENABLED="true" # Enable caching

# Configure Custom Properties
# Note: There is a lock file that allows you to configure properties dynamically
export HELICONE_PROPERTY_<PROPERTY_NAME>="<PROPERTY_VALUE>" # Set custom properties

bash -c "$(curl -fsSL https://raw.githubusercontent.com/Helicone/helicone/main/mitmproxy.sh)" -s start
echo "sk-<HELICONE_KEY>" > ~/.helicone/api_key # Alternatively, you can export HELICONE_API_KEY to your path before starting the proxy to bypass this step.

export REQUESTS_CA_BUNDLE=/etc/ssl/certs/ca-certificates.crt # This is required for HTTPS requests to work properly.
# Your python script here...

​

Stopping the Proxy

bash -c "$(curl -fsSL https://raw.githubusercontent.com/Helicone/helicone/main/mitmproxy.sh)" -s stop

​

Viewing Logs

bash -c "$(curl -fsSL https://raw.githubusercontent.com/Helicone/helicone/main/mitmproxy.sh)" -s tail

​

Installation (Mac)

To install on a Mac, follow the same instructions as for Ubuntu/Debian, but use the mitmproxy_mac.sh script instead. You can find this script here: https://raw.githubusercontent.com/Helicone/helicone/main/mitmproxy_mac.sh

​

Configuring the Dynamic Helicone Headers

You can configure custom properties using a lock file located at ~/.helicone/custom_properties.json. These properties will be applied to all requests.

{
  "<PROPERTY_NAME>": "<PROPERTY_VALUE>"
}

To dynamically configure the proxy, you can install our python package.

pip install helicone

Then, use the following code:

from helicone.lock import HeliconeLockManager
HeliconeLockManager.write_custom_property("job_id", "1")
'''
Other Methods:
 - write_custom_property(property_name: str, value: str) // Write a custom property
 - remove_custom_property(property_name: str) // Remove a custom property
 - clear_all_properties() // Clear all properties
'''

This enables you to manipulate properties dynamically, making your workflow more flexible and efficient.