Go to the “Knowledge” tab. This is where you’ll set up everything and add your new embeddings. Make sure you have selected an environment in the top select, or the button will be blocked until you do so.
Prerequisite
To add embeddings to your knowledge environment, you will need to do two things first:
- a knowledge environment.
- an AI environment with a usable embedding model matching your vector database dimensions.
The embeddings model is set in Settings → AI → Default Environments for AI → Embeddings (you can also override it per knowledge environment). Your setup should look something like this:
The Dimensions must match the size of your vector database index. With a Custom (OpenAI-compatible) model, AI Engine can’t know the size for sure, so the field stays editable: check your model’s documentation and enter it yourself. If no embeddings model is set at all, AI Engine shows a warning in the Knowledge settings and in the chatbot editor, since Knowledge can’t search without one.
Manual
The simplest way to create a new embedding is to do it manually. Click the green “+” button or the “Create New” option under Build Knowledge, and a new modal will appear.
From there, set a title and fill in the content. The behavior and type fields will be locked to “context” and “manual”, since that’s the mode you’re currently using.
The Title is not part of the embedding itself. It’s only there to help you manage and organize your entries more easily. If you want the title to be included in the actual embedded data, you’ll need to also add it manually inside the Content field.
Embeddings based on Posts
AI Engine includes built-in tools to help you generate embeddings based on your WordPress posts. This includes automatic synchronization when posts are modified and automatic deletion of embeddings when posts are deleted.
You can also choose a specific post type to use for embedding sync. This gives you flexibility, you can create a custom post type (CPT) specifically dedicated to embeddings. This way, you can design advanced behaviors tailored to your content structure, using embeddings that are cleanly separated from your regular content. This setup is especially useful if you’re aiming to create dynamic interactions or content-aware chatbots.
In the right-hand section, under Environment, open the Sync tab and click “Enable Sync”. Then confirm by clicking “Use Current Environment” to sync the environment currently open in the Knowledge tab.
You’re done! All the post types listed here will now be synced automatically. If you encounter any issues with the sync, please check the documentation for troubleshooting steps.
You can also do that manually, without the sync, by using the “Push All” button and selecting a post type; this will go through all the posts from that post type and create embeddings based on their content. If the sync is not enabled, then it’s just like manual embeddings; if the sync is enabled, they will also auto-update.
When it’s done, AI Engine shows a short report: how many posts were added, how many were already up to date, and how many were skipped, with the reason for each skipped post (for example, a post that is ignored, filtered out by the language or category settings, or that has no text). If many posts are skipped with “No text was found” and your pages are made with a page builder (Breakdance, Oxygen, Bricks, Elementor…), turn on Shortcodes: Resolve in Settings → Others → Advanced, then push again.
Each post becomes one embedding. AI Engine keeps at least the first 16,384 characters of each post (more if your Context Max Length is set higher), so keep your most important information near the top of very long posts.
Advanced Sync Behaviors
More options are in Settings → Knowledge → Sync Settings → Local Content. Force Recreate re-creates the embeddings every time a sync runs, even when the post content, the model and the dimensions haven’t changed. Rewrite Content is different: it asks your default AI model to shorten and improve the content before it is embedded (this costs one extra AI query per post).
This is useful if some of your content comes from custom fields (like ACF) or third-party builders, since that data might not be part of the main post body. With Force Recreate on, the data added by the Content Parser addon is refreshed on every sync.
Embeddings based on PDF
To create embeddings from a PDF file, go to the Build Knowledge section on the right side of the Knowledge tab and click the “Upload PDF” button. You can then follow the step-by-step process to create embeddings.
This approach lets you do everything on WordPress, but the tradeoff is you have to chunk the files, so it requires a bit of preparation.
If you prefer to just dump your files and not worry about it, you can rely on a Vector Store, which supports direct file upload, so you can just “throw” your PDF in there instead. But for this you will rely on the OpenAI Vector Store, which is a third party; you can read here to set it up: Vector Store.
Embeddings based on URLs
You can create embeddings from URLs, whether they are internal or external. The system will scrape the page and extract its readable text content.
Keep in mind that the quality and structure of the extracted text depend heavily on the original page formatting. The content is processed as raw text, so HTML elements such as hyperlinks are not preserved unless the full URL is explicitly written in the text.
To create one, click the “+” or “Create New” button to open the Add Embedding popup, then change the Type to “Remote URL.” Add the corresponding URL and Title for the content you want to embed.
If you enable Expert Mode, you can create bulk URL embeddings. This allows you to paste multiple URLs separated by line breaks, and the system will process them all at once.
By default, Remote URL embeddings are synced every 24 hours, meaning the content is refreshed from the same URL daily. You can change this frequency by going to Settings → Knowledge → Sync Settings → Remote URLs → Sync Interval.