Dynamic CPT’s AI Assistant turns plain prompts into ready-to-use schemas for queries, listings, relations and sample data. Create something new, modify an existing setup, or repair a broken schema, all in a few clicks.

AI Assistant #

Let’s see the controls here.

Mode: Create New #

Example pompt: Create a non-hierarchical taxonomy called “Skill Tags” for the “Job” post type, similar to tags (no parent-child). Fields: name, slug. This should support multi-select so a job can have multiple skill tags.

You can also choose to add data filtering/search (Queries), a frontend listing page (Listings), or a connection with the post type (Relations) before generating the schema.

Your AI powered taxonomy is ready to apply.

Apply Schema
This means: when you click “Apply Schema (2)”, the schema you generated gets applied to your site, and before applying it, the tool automatically saves a snapshot (backup) of the current state.

The “(2)” indicates 2 items/changes are being applied in this schema.

If anything goes wrong after applying, you can use that snapshot to roll back with one click. This means you can instantly undo the changes and restore things to how they were before.

We’ve applied the schema. Here’s a quick demo of the process in action:

Mode: Modify Existing #

Context: Agents
Target Field Group: Any

Example pompt: Add a phone number field (text) to Agents. Scroll down and click Generate Schema.

Here is the schema. The phone number meta field is ready for the post type “agent”.

After applying, this is how it will look when I go to the Agents post type.

Mode: Fix Schema #

1. Example pompt: Fix the JSON syntax errors, and change the reserved slug “post” to “agent-profile”.
2. Schema:

{
  "post_types": [
    {
      "slug": "post",
      "label": "Agents",
      "supports": ["title", "editor", "thumbnail"
    }
  ],
  "fields": [
    {
      "name": "phone_number",
      "type": "text"
      "label": "Phone Number"
    }
  ]
}

3. Click Generate Schema.

The errors are now fixed, and you can apply the schema as before if you want.