Introduction to Prompt Sentinel

Prompt Sentinel is a Python library designed to protect sensitive data during interactions with language models (LLMs). It automatically detects and sanitizes confidential information—such as passwords, tokens, and secrets—before sending input to the LLM, reducing the risk of accidental exposure. Once a response is received, the original values are seamlessly restored. This makes the masking process transparent to the client, enabling smooth and uninterrupted interactions, including function calling.

Prompt Sentinel focuses on simplicity—it works out of the box without requiring complex proxy-based LLM deployments or infrastructure changes. It offers flexible detection options, ranging from regular expressions to trusted local or external LLMs, enabling accurate identification of sensitive data such as personally identifiable information (PII) and secrets. Both detection and restoration processes are easy to set up and highly configurable to suit a wide range of applications. To avoid redundant detection calls—especially when relying on LLM-based detection—Prompt Sentinel includes a smart caching mechanism that boosts efficiency and significantly reduces overhead.