How to Use AI to Boost Gravity Forms Moderation (with GC OpenAI)

1. Step 1 — Install and Activate GC OpenAI
2. Step 2 — Create a GC OpenAI Feed
3. Step 3 — Set Up Moderations Endpoint
4. Step 4 — Test Moderation
5. You’re all set!

Gravity Connect: OpenAI empowers you to take advantage of OpenAI’s moderation models to review and flag form submissions based on OpenAI’s Usage Policies. It takes the brunt of internet trolls and disrespectful users so that you and your team don’t have to.

But it’s not just about defense—it also helps you move faster. For things like product reviews or job postings, it can automatically screen submissions so you can publish with confidence, without second-guessing what might slip through. Especially if you already want to use GC OpenAI for generating texts, images, and/or audio directly from Gravity Forms, this is a nice little cherry on top.

Let’s see how that’s done using a “Leave a Review” form as an example!

Step 1 — Install and Activate GC OpenAI

GC OpenAI is available through Gravity Connect. You can find the download under Account > Downloads > Gravity Connect.

Next you’ll need to set up the OpenAI API, but it’s magically simple to do so.

Step 2 — Create a GC OpenAI Feed

Head to the form you want to set up moderation for. In the Settings menu, click on OpenAI to get to the GC OpenAI feeds screen. From there, click on Add.

Step 3 — Set Up Moderations Endpoint

Under General Settings, select Moderations as your OpenAI Endpoint.

Moving onto Moderations, you can choose which moderation Model to use. Currently, GC OpenAI offers two endpoints that communicate with the exact same version of the `text-moderation` model, so you can leave the default selected and move on.

Past Model down to Input, use merge tags to decide which fields will be sent to the model. Note that this model does moderation, not generation—only send content that you want moderated. The default is the {all_fields} merge tag.

Next, use Behavior to choose what GC OpenAI should do with the entry if it gets flagged by moderation.

How it works: Submissions are scored from 0 to 1 in multiple categories. The higher the score, the nastier the submission is. High-scorers get flagged, GC OpenAI generates a report and can then either:

• Do nothing: Allow the form to be submitted and just have the OpenAI report for reference in the entry meta. (Screenshot below.)
• Mark entry as spam: Allow the form to be submitted, but mark entry as spam. This way you can keep it for additional review and recordkeeping while not running feeds or notifications for it.
• Prevent submission by showing validation error: Block form submission and communicate with the user via validation message. You can personalize this message to fit your use case. The default is This submission violates our content policy.

To top it all off, you can use conditional logic to only run the feed when necessary. For example, I have a handful of long-time customers that I trust, so I don’t need to spend OpenAI tokens to moderate their reviews. I set up a custom meta field in their customer profiles to tag them as trusted. Then, added a hidden field and set the default value to {user:customer_type} to be able to use it in the feed’s conditional logic, like so:

Lastly, under Advanced Settings: Moderations, you can adjust the Timeout setting, aka the number of seconds OpenAI should wait to respond. You’ll typically only need to tweak this if you’re troubleshooting. We recommend giving it a minimum of 5 seconds, which is the default.

Step 4 — Test Moderation

Do your worst!

You’re all set!

Your forms are now protected with the magic of AI. If you have any questions about the setup or OpenAI, let me know in the comments!