Allowing ChatGPT Access via Deep Packet Inspection Allowlist

Overview

This article explains how to configure Endpoint Protector to allow access to ChatGPT by adding the openai.com domain to a Deep Packet Inspection (DPI) allowlist. In environments where DPI is enabled, web-based AI services like ChatGPT may be blocked by default because security policies restrict access to certain domains.

tip

Since EPP Client release 2511, Endpoint Protector offers wide native Data Loss Prevention coverage for AI/LLM providers, including ChatGPT, Microsoft Copilot, Google Gemini, DeepSeek, X Grok, Claude, Meta AI, Perplexity, and Cursor. Instead of only allowing or blocking access at the domain level, you can monitor and control the content exchanged in AI prompts. See Content Aware Protection for AI interactions for the full list of supported AI platforms and configuration examples.

Instructions

1. In the Endpoint Protector Console, navigate to Denylists and Allowlists > Allowlists > Deep Packet Inspection.
2. Create a new allowlist dictionary for allowed domains and add the domain using wildcards: *openai.com*
3. Once the allowlist dictionary is created, include it in the relevant Content Aware Policy.
4. Edit the Content Aware Policy.
5. Expand the Policy Allowlist section.
6. Navigate to Deep Packet Inspection and select the dictionary you created.
7. Save the policy.