If you’re tired of copying code back and forth between your AI editor and WordPress, Novamira is here to remove that friction. It puts an AI bridge inside your WordPress install so your agent can work with the site, not around it.
We approached this video with minimal prep, no sponsorship, no affiliate links, and no attempt to make the product look perfect. The core setup worked, the early results were impressive, and the experimental parts felt experimental. That’s the best way to start.
Key Takeaways
- Novamira brings AI into WordPress through MCP, which lets the agent work directly inside the site instead of relying on back-and-forth copy and paste.
- Novamira handled our tests extremely well, whether it was adding forms to the WordPress block editor, editing pages with page builders, or creating products within WooCommerce.
- The free version connects and works, while Pro adds memory (among other things), which is a useful upgrade for repeated site work and longer sessions.
- Novamira Visual shows strong potential, and even as an experimental feature, it adds a promising visual workflow to the plugin.
- The best way to use Novamira is first on staging or development, which keeps testing safe and gives you room to explore its AI features with confidence.
Novamira—First Look Video
What Novamira Is Trying to Fix
The pitch behind Novamira is easy to understand once we strip away the AI buzz. Without a tool like this, asking an AI to help with WordPress often means a lot of manual back-and-forth. We copy code out of Cursor or another AI client, paste it into WordPress, check the result, go back to the AI, explain what changed, and then repeat.
Novamira cuts into that loop by letting the AI connect to WordPress through MCP. Instead of guessing, the agent can inspect the site, read data, execute PHP, work with files, and respond to what it finds. That matters because WordPress isn’t one flat codebase. It’s the database, the theme, the active plugins, the settings, the post meta, the page builder content, and everything else that are glued together.
At a simple level, the setup looks like this:
- Your WordPress site
- The Novamira plugin inside that site
- Your AI client, such as Cursor
That middle layer is the whole point. The AI doesn’t work from pasted snippets alone. It’s connected to the actual environment. According to the product description, that includes PHP execution, database queries, file edits, and WP-CLI access, along with plugin-specific abilities where supported.
This functionality also helps explain why the product has drawn attention from WordPress professionals. If it works, we spend less time switching tabs and less time translating the same context over and over. The AI can test, inspect, adjust, and try again inside one conversation.
What the Website Gets Right, and What’s Missing
Documentation, Features, and Trust Signals
Before installing anything, we wanted the usual trust pages. That means the pages that help a new customer figure out who built the product, how active it is, and whether the documentation is solid.
The first thing that stood out was a missing About page. We’d like to see one. Names, faces, product background, and a clearer explanation of the connection to the Dynamic.ooo team would help a lot. There is a hint of that relationship on the site, but it’s easy to miss.
Documentation, on the other hand, is great! It was easy to follow with video sprinkled in and felt organized. The changelog was also the kind of thing we like to see: version numbers, release dates, and real details about what changed. For a product moving this fast, that’s important.
The feature pages also did a good job of speaking to the intended audience. If you’re a WordPress builder using Elementor, Bricks, Divi, WooCommerce, or developer-focused workflows, the technical examples make sense. Things like regenerating thumbnails or converting a page to Elementor V4 atomic widgets aren’t broad marketing fluff. They’re concrete use cases.
One area that could be clearer is the WordPress block editor. There was plenty of emphasis on page builder workflows, especially Elementor and Bricks, but less obvious detail on what the plugin can and can’t do with Gutenberg. That doesn’t mean the support isn’t there. It means the website could surface it better.
We found the Novamira does just fine with the WordPress block editor (Gutenberg) and demonstrated that in the video.
For anyone looking at the product cold, the Novamira website explains the core idea well. It could still do more to tell the story of the team behind it.
Free vs. Pro Without the Confusion
The free vs. pro conversation around Novamira doesn’t need to be complicated. The most significant difference is memory.
With the free version, the AI can still connect and work, but it doesn’t retain the same level of persistent recall. That means more repeated prompting and, in many cases, more token usage because the agent has to rebuild context from scratch more often.
Pro unlocks memory—it remembers your stack. That sounds small until you picture daily use. If the agent remembers your site setup, your preferred tools, your naming conventions, and the work you were doing before, the whole workflow gets more efficient. It also gets less annoying.
For testing, the free version makes sense. For regular use, memory seems to be one of the main reasons Pro exists at all.
Setting Up a Safe Test Environment
Keep AI abilities in staging or development. Leave them off in production.
That best practice was clear and important. Novamira can stay installed on a live site, but the AI abilities should remain off there. The active work should happen on staging or in a development environment, and then you can push changes over in the normal way.
The setup started on a test site with Elementor, Elementor Pro, Fluent Forms, and WooCommerce already installed. One early surprise was that Novamira wasn’t pulled from the WordPress plugin repo during the test. The plugin was installed by manually uploading the zip downloaded from the product site.
That may not be a problem once you know it, but it’s the kind of thing a first-time user notices right away. We first searched the WordPress repo. That’s what many users will do.
Install, Node.js, and AI Abilities
Once the plugin was installed, the next step was turning on AI abilities. Those are disabled by default, which is the right call. Novamira is giving an AI agent access to powerful actions. Such actions shouldn’t be on by accident.
After activation, the admin area exposed a few important pieces. There was a configuration area for enabling AI abilities, an abilities hub for managing what gets exposed to agents, a context section, a skills section, memory settings in pro, and a visual feature marked experimental.
A Windows setup also required Node.js during the test. The docs walked through it, and that part was straightforward. Node.js is needed to run the MCP remote proxy when connecting AI clients to the WordPress site. Once that was installed and verified, the rest of the connection flow moved forward.
It was also encouraging to see plugin-aware behavior show up early. Novamira detected installed tools and exposed specializations around them. In the test environment, that included Elementor and WooCommerce, and the expectation was that Fluent Forms support was available as well.
Connecting Cursor With an Application Password
Authentication relied on a WordPress application password. That process was smooth.
Inside the Novamira settings, it generated the application password, embedded the connection text needed for the AI client, and guided the next step. In Cursor, that meant pasting in the generated prompt and letting the client configure the MCP connection.
Verification was simple enough. In Cursor, the connection showed as active under the MCP tools area, and a quick test prompt confirmed it. The first real proof was a command to list the plugins installed on the WordPress site. It returned the installed plugins, which was the first sign that the site and AI client were talking properly.
From there, the test moved from setup to the fun part.
What Worked in the Core Product
The fastest way to judge a tool like this is to stop reading about it and give it real prompts. That’s what happened next.
We didn’t build custom skills first. We didn’t spend time polishing rules. The test started with basic, plain-English requests to see whether Novamira could do useful WordPress work with supported plugins.
Fluent Forms and WooCommerce in Action
The Fluent Forms test was simple. A new form named “Test Form” was created through a prompt. It appeared in WordPress on refresh.
Then the form was updated with two more prompts. One added a simple text field and a website URL field. Another changed the submit button and the form’s brand color treatment to red. Again, the changes showed up after a refresh.
On paper, that may sound like a small win. We could log into WordPress and do those things by hand. The real point is that we didn’t have to leave the AI workspace to do it. That’s where the value becomes clear.
WooCommerce made that even clearer. A prompt created a simple product with a regular price of $99, a sale price of $59, and a sale end date of August 1, 2026. The product details were filled in correctly.
A second prompt created another version of the product that was not on sale. Then both products were linked together as upsells and cross-sells. Thereafter, another prompt changed both prices to $89. That last change also removed the sale setup because the prompt wasn’t specific enough. That was a useful reminder that the agent still does what we asked, not what we meant.
For store work, this is where things get interesting fast. One product is nothing. Ten products are different. Hundreds of products with shared rules, linked groups, subscription logic, inventory settings, sale windows, images, and tags—now we’re talking about real saved time.
The same applies to forms. Adding one field isn’t the story. Generating an entire multi-step survey, using brand colors, and doing it while we keep working elsewhere—that’s the story.
Gutenberg and the WordPress Block Editor
The block editor tests were one of the more pleasant surprises.
A prompt created a new post titled “Novamira Is Great” and applied a tag of AI. That worked. Then a page named “Novamira” was created and used for a more useful test.
The page started empty. A prompt asked for a heading titled “Sign Up for Our Newsletter” and requested that the previously created Fluent Form named “Test Form” be added underneath it. No form ID was provided. Only the form name.
Novamira handled it. It added the heading and inserted the form block content needed to render the form, and the front end displayed the expected result.
After that, another prompt changed the heading to title case and changed the form button color to blue. That update also worked.
This change matters for a few reasons. First, it indicates that Novamira isn’t limited to developer tasks or page builder workflows. Second, it suggests the block editor support is stronger than the website initially makes obvious. Third, it shows that the agent can work with block markup in a way that feels natural from the user’s side.
There was also a background utility page inside Novamira related to validating and serializing Gutenberg blocks. That detail won’t matter to every user, but it does suggest that the block editor support team is handling it behind the scenes in a careful way.
Why Skills, Context, and Memory Matter More Than the Demos
The demos are engaging. The long-term workflow is where Novamira either earns a spot or doesn’t.
Skills are part of that. During the test, they were described as shorthand commands that let us avoid typing the full instruction every time. We could build a skill that simplifies “list the plugins installed on this WordPress site in alphabetical order by status” to just “plugins” or “list my plugins.”
That sounds small until we consider daily use. The more repeatable tasks we offload, such as product setup, WooCommerce rules, form structures, and page skeletons, the more helpful those reusable prompts become.
Context is another big piece. Novamira includes a place to store site goals, brand voice, audience notes, naming conventions, and similar instructions. That means the agent can start with useful background every time instead of asking us to restate basics. The newer Context page in Novamira 1.7.0 pushes that idea even further.
Then there’s memory. This feature is the part that can make Pro worth it for serious use. If the AI remembers what we were building, what plugins we’re using, and what patterns we prefer, the session becomes more practical and less repetitive.
All of this ties back to one core benefit: less context switching.
That’s the strongest case for Novamira. We can stay in our AI editor, keep working on plugin code, content structure, performance fixes, or whatever else is in front of us, and let the AI handle the WordPress dashboard work in parallel. Type the request. Or say it if you’re using voice tools. Then come back when the job is done.
How Novamira Visual Worked
Novamira Visual is the feature that will grab attention because it promises something easy to imagine. Instead of the AI working off-screen, it opens pages and editors in a browser workspace so we can watch the work happen live.
That was the expectation going in, and the docs backed it up. The product also labels Visual as experimental, which turned out to be an important label.
Pairing the Visual Sandbox
The visual setup used a different prompt than the core connection. After that, the AI client showed a visual sandbox MCP entry, which was a positive sign.
The browser-based pairing was less smooth. During the test, there was some confusion before the pairing code appeared. The AI client mentioned being logged into WordPress, leaving the screen open, and being on the same computer. Then the code appeared and the connection completed.
That part felt rough. Not broken forever, but rough enough that a new user could pause and wonder if they missed something. A little more onboarding or clearer status feedback inside the visual interface would help.
Once paired, the expectation was simple. Open a page, provide a prompt, and watch the edits happen live.
Live Gutenberg Edits Looked the Part
After the visual connection was sorted out, Gutenberg behaved much more like we expected.
A button was removed from a page through a prompt, and the change happened live. Then a more complete test asked for an “FAQs” heading and five accordion items titled “Question One,” “Question Two,” and so on, with a blue background, light text, and accessible color contrast.
The page built in front of us, with the heading appearing first, followed by the accordions. The heading appeared first, then the accordions. On the AI side, there was also a useful summary of the structure and an accessibility report tied to the color choices. It even reported WCAG compliance for the selected colors.
That live feedback changed the feel of the tool. Instead of prompting and refreshing manually, we could see that the agent was doing work in the page editor while we watched. That’s the version of Visual most users will want.
Elementor Showed Promise but Also Rough Edges
Elementor was more mixed.
The first attempt asked Novamira Visual to create a new page and open it in the Elementor editor. The page was created, but it didn’t open itself in the editor. Opening it manually fixed that part.
From there, a prompt asked for a full atomic form element. Novamira did add elements, and the page refreshed while it worked, but the result wasn’t a full form in the way we expected. It got part of the job done, including a blue button and hover styling, but not the full target.
That wasn’t a problem because the feature is labeled experimental. It was a reminder to keep expectations aligned with that label.
A second Elementor test went better. All widgets were removed, then a heading for an FAQ section and a five-item accordion were added with the same blue-and-light-text accessibility request used in Gutenberg. That result landed much closer to the target, and the AI again returned a small accessibility summary.
So where does that leave Novamira Visual? Gutenberg looked stronger in this session. Elementor showed promise, but it also had the kind of friction we’d expect from an experimental feature still being pushed forward.
What We’d Like to See Improved
The biggest website improvement is still the missing About page. When a product is asking for deep trust, direct WordPress access, AI permissions, file edits, or PHP execution, people want to know who built it. That information shouldn’t stay hidden in hints and side references.
We’d also like more obvious messaging around block editor support. The product proved it can do useful Gutenberg work, but that wasn’t as clear from the website as it could have been.
On the product side, Novamira Visual needs a tighter pairing experience. The live editing concept is effective The early confusion during connection creates avoidable friction.
None of those points cancel what worked. They matter because the core idea is strong enough that the rough edges are worth fixing.
Our Take on Novamira
Novamira is one of the more compelling WordPress AI products we’ve looked at because it solves a real workflow problem. The appeal isn’t that it can add one field to a form or create one product. The appeal is that it lets us keep working while the AI handles WordPress tasks in the background, with access to the actual site context.
The core product left a strong first impression. The experimental visual layer had bumps, but it also showed enough to keep our attention. For agencies, developers, store owners, and serious WordPress DIYers, context is the feature to watch. That’s where the experience starts feeling less like a demo and more like a working system.
If you like this kind of direct, unsponsored product coverage, the InfluenceWP Partner Membership is where product teams can request the same kind of first-look review.
Frequently Asked Questions About Novamira WordPress Plugin
What Does Novamira Do for WordPress?
Novamira connects an AI client to a WordPress site through MCP, so the agent can inspect the site, run PHP, query the database, edit files, and handle plugin-aware actions inside the real environment. It helps reduce the copy-paste loop between an AI editor and WordPress.
Does Novamira Work With Gutenberg?
Yes. The article demonstrates how it works with the WordPress block editor by creating a page, adding a heading, inserting a Fluent Forms block, and changing the button color. That makes Gutenberg one of the clearest strengths in the review.
What is the Difference Between the Free and Pro Versions?
The free version connects and works well, while Pro adds memory. That makes Pro especially helpful for repeated site work, ongoing projects, and workflows where the agent needs to remember setup details.
Should Novamira Be Used on a Live Site?
The best approach is to use Novamira on staging or development first. That gives you a safer space to test ideas while keeping production stable.
How Well Does Novamira Visual Work?
Novamira Visual looks promising and adds a useful visual layer to the plugin. It is still experimental, but the live editing experience already shows real potential, especially in Gutenberg.