Prompt Tracing in Python

An example Python app demonstrating how to integrate Pangea's Secure Audit Log service to maintain an audit log of prompts being sent to LLMs.

Prerequisites

• Python v3.12 or greater.
• pip v24.2 or uv v0.4.29.
• A Pangea account with Secure Audit Log enabled with the AI Audit Log Schema Config.
• An OpenAI API key.

Setup

git clone https://github.com/pangeacyber/python-prompt-tracing.git
cd python-prompt-tracing

If using pip:

python -m venv .venv
source .venv/bin/activate
pip install .

Or, if using uv:

uv sync
source .venv/bin/activate

Then the app can be executed with:

python prompt_tracing.py "What is MFA?"

Usage

Usage: prompt_tracing.py [OPTIONS] PROMPT

Options:
  --model TEXT            OpenAI model.  [default: gpt-4o-mini; required]
  --audit-token TEXT      Pangea Secure Audit Log API token. May also be set
                          via the `PANGEA_AUDIT_TOKEN` environment variable.
                          [required]
  --audit-config-id TEXT  Pangea Secure Audit Log configuration ID. May also
                          be set via the `PANGEA_AUDIT_CONFIG_ID` environment
                          variable.
  --pangea-domain TEXT    Pangea API domain. May also be set via the
                          `PANGEA_DOMAIN` environment variable.  [default:
                          aws.us.pangea.cloud; required]
  --openai-api-key TEXT   OpenAI API key. May also be set via the
                          `OPENAI_API_KEY` environment variable.  [required]
  --help                  Show this message and exit.

Example

This does not modify the input or output so it's transparent to the LLM and end user.

Audit logs can be viewed at the Secure Audit Log Viewer.