ChatGPT Plugin Policies

Here are the key guidelines set by OpenAI for developers who are developing ChatGPT Plugins:

The functionality of the exposed API must be clearly outlined in the plugin manifest file.
Unnecessary, deceptive, or irrelevant terminologies or instructions should not be included in the plugin manifest, OpenAPI endpoint descriptions, or plugin response messages. This also covers instructions that seek to alter model behavior or discourage the use of other plugins.
The use of ChatGPT plugins to bypass or tamper with OpenAI’s safety mechanisms is not permitted.
Using plugins to simulate human-like responses or automate conversations with actual individuals through pre-programmed messages is prohibited.
If a plugin is used to disseminate personal communications or content generated by ChatGPT (such as emails, messages, or other types of content), it should clearly state that the content was generated by an AI.

Just like other usage policies, OpenAI anticipates alterations to the current ChatGPT Plugin Policies as they gather more information about the potential use and misuse of plugins.

Submitting Plugin for Review

You can expect to hear back about a plugin you submit for review ~7 days after you submit the plugin.

OpenAI is currently evaluating new ChatGPT Plugins on a continuous basis. You can make use of the plugin submission bot to submit your plugin for review. However, you must be signed in to access the plugin submission bot.

To monitor the progress of your plugin submission, ensure that you are logged in and navigate to the “Help” option situated at the top right corner of the page. Here, under the “Messages” section, you will find your plugin submission. You will receive notifications as your plugin advances through different stages of the review process.