[00:00:20] Nathan Wrigley: Hello there and welcome once again to the WP Builds podcast. You've reached episode number 422 entitled, Streamlining WordPress Agency Workflows: Zach Hendershot introduces Miruni. It was published on Thursday, the 22nd of May, 2025. My name's Nathan Wrigley and a few bits and pieces just before we begin.
The first thing to mention is a big thank you to everybody who participated, organized, turned up for, did anything with, the Page Builder Summit, which happened last week. Really appreciate it. We had many, many satisfied people. Lots of presentations, nearly 40, I think it was by the final count. Lots of networking. So really appreciate that. I hope that you enjoyed that, and hopefully we'll be bringing version nine of that event to you soon.
The other thing to mention is that if you are enjoying the WP Builds podcast, why not subscribe? Head to wpbuilds.com/subscribe. And over there you'll be able to find many different ways that you can subscribe, be it on YouTube, be it following us on Twitter or X whatever platform. But probably the best way that I can recommend is to sign up for our email newsletter, which is on that page, but also subscribe to us in your podcast player of choice search for WP Builds. That truly is the best open web way of keeping up to date with everything that we do.
Also, if you are a product owner, perhaps you've got a service, a plugin, a block, a theme. Maybe you're into hosting in the WordPress space, and you would like to get in front of a WordPress specific audience. Well, we have exactly that. Head to wpbuilds.com/advertise to find out more. A bit like these three fine companies did.
The WP Builds podcast is brought to you today by GoDaddy Pro. GoDaddy Pro, the home of managed WordPress hosting that includes free domain, SSL and 24 7 support. Bundle that with The Hub by GoDaddy Pro to unlock more free benefits to manage multiple sites in one place, invoice clients and get 30% off new purchases. Find out more at go.me/wpbuilds.
The lights are kept on this week also by Bluehost. Bluehost, redefine your web hosting experience with Bluehost Cloud. Managed WordPress hosting that comes with lightning fast websites, 100% network uptime, and 24 7 priority support. With Bluehost Cloud the possibilities out of this world, experience it today at bluehost.com/cloud.
And Omnisend are helping us with our podcast this week, Omnisend, do you sell your stuff online? Then meet Omnisend. Yes, that Omnisend. The email and SMS tool that helps you make 73 bucks for every dollar spent. The one that's so good, it's almost boring. Hate the excitement of rollercoaster sales? Prefer a steady line going up? Try Omnisend today at omnissend.com.
And sincere thanks, go to GoDaddy Pro. Bluehost and Omnisend for their support of the WP Builds podcast. Podcasts like this simply cannot happen without the support of fine companies like that.
Okay. What have we got for you today? Well, I'm joined today by Zach Hendershot.
This is a really interesting, and I've got to say fairly new, a unique product as far as I can see. It's called Miruni, and it's an AI product, and at this moment you're thinking, oh, yawn, AI, everybody's talking about AI. But this is really quite different.
The idea of this, it will help you if you have client websites, and those clients are constantly making changes. The idea here is that they upload a screenshot and make modifications on that screenshot, and AI figures out what the changes will be so that you just have to click publish. Now, let that sink in for a moment. The idea is really interesting and this podcast is all about that, and I hope that you enjoy it.
I am joined on the podcast by Zach Hendershot. Hello, Zach?
[00:04:31] Zach Hendershot: Hello, how are you?
[00:04:32] Nathan Wrigley: Good. Very nice to meet you. Zach's gonna be talking to us today about, actually, I don't know if it's a brand new thing or not, in the tiny little conversation that we had before we hit record.
You alluded to the fact that maybe something of this nature has been around for a little while. We'll get into that in a moment. It's called Miruni. And, maybe you wanna pause the podcast now and go and Google that. It, the spelling of that is, M-I-R-U-N-I Or make it simpler. Just go to Miruni.io.
That's gonna be a nice, easy thing to do. So Zach, just before we begin, just tell us a little bit about who you are, where you're from. Let's, we'll just keep it fairly brief and, related to WordPress and SaaS and things like that.
[00:05:15] Zach Hendershot: Awesome. Yeah. I'm, I'm Zach Hendershot. I, I've been an engineer in building WordPress sites and building SaaS businesses for many years At this point, I, I'm an engineer by trade, but I've, I. I've been in the, product and design space and user experience space for, many years, and I'm Pa and I've run many agencies as well in the past, started my own, had leadership roles in a variety of them.
So I've been passionate about building great experiences and ultimately delivering like world-class web experiences, which is a lot of the orgs and story of, what we're building here with Miruni.
[00:05:46] Nathan Wrigley: And I guess that you've also been slightly frustrated with some of the tasks that were involved in running an agency, because basically that's what Miruni is all about, isn't it? It's about giving the agencies that are building WordPress websites and by extension their clients. A nice, simple, easy interface so that they can communicate with each other and cut out wasted time and things like that.
Now, there's my elevator pitch of what it is, but over to you. Tell us, in the elevator pitch style, what it is that Miruni is and does and how it works.
[00:06:20] Zach Hendershot: That is a good elevator pitch. Yeah. We're, focused on automating away a lot of the manual boring work that, so many of us spend hundreds of hours, doing inside of, WordPress agencies and building websites. so yeah, Marney is a, an AI tool that automates edit requests come from, coming from your customers and your clients.
so we have an AI engine that can take the request and automate the resolution of that request. Implement what's being requested, automatically in, a lot of cases, in, in, our feeling, 80% of the requests that come in can just be automated for you. and relegated to a push button resolution of the issue versus, eight to 10 clicks per issue to go and find it, update it, preview it, publish it, the, whole nine yards.
So we're on a mission to, to rip all that boring stuff outta your day to day.
[00:07:08] Nathan Wrigley: So a lot of the people listening to this podcast are probably doing something else. They're probably, I don't know, mowing the lawn or out on a run or something like that. So in order to drill that home, I'm gonna re redescribe it because I watched a video, I. Of what your product does. You sent me a link for a video and I will put that into the show notes 'cause it's pretty cool.
So imagine the scenario where you, dear agency owner, have a client and the client wishes to make a modest change. Now, the typical setup, you may have this perfect system, right? you may have spent absolutely ages getting this exact perfect system, right? But if you haven't, this is what Miruni does, the client.
Basically it toggles something in the browser, which enables like a screenshot piece of software and they, just drag and highlight the bit on their webpage that they want to change. So let's imagine there's a paragraph and they just wanna swap a couple of words out. they were becomes, I am, or something like that.
Then they describe what it is that they want to do and fire it off, click go. And from their point of view as a client, they're dumb. Largely that's, their interaction over it, then gets fired off by Miruni you as the agency owner. Get a notification that something needs to be changed, but here's the cool bit.
This is the million dollar bit. You don't need to do the change. The change is interpreted by an ai, and all you need to do is go into something which looks like version control. So there's what it used to look like. Here's what the AI thinks it should be updated to, and the human goes. Okay. Is that right?
Yeah, that's right. Click go if you like, save, publish, whatever, and it's done. So the magic bit in the middle there is the AI has somehow figured out what page the client was on, which paragraph the client was in, which word the client wanted to change. And it just suggests it. And you just say, yeah, that's it.
And then, you're done. Look, that's really clever, Zach.
[00:09:21] Zach Hendershot: Yeah, no, that, you, you described it perfectly, right? Like we're to give a little bit more background on what, around, what we're doing and how we're, doing that magic. we're collecting the screenshot, as you said. We're collecting a lot of background information that goes to, things like, what the web browser or the screen resolution, things that might be relevant to do the request in the most accurate way possible.
That goes to an LLM, that LLM has, access to content. The posts, comments, pages, theme information, CSS, information about the theme, including configuration information that may be configuring the page to present, graphs or screenshots or, different bits of, visual information.
All that is available to the LLM to make a direct, hopefully very targeted change based on the, simple request that came in and the context it has, and it gives the website admin a. Very simple tool to see the before and after, as you said, and you just click and apply the button and, we manage the workflow.
We also, c communicate that change to the client as well. So it re reduces some of the burden of lots of manual emails and slack messages and whatever. and, and makes it a, very uneventful, and hopefully very fast, activity for the website admin to, to get, the request done fast, and deliver great
[00:10:40] Nathan Wrigley: Yeah. So fast is the word there, right? Because from the client's point of view. You invoke this screen recorder thing, which, there's, like almost no options. You just click the button, drag the area and type what you want. But from the point of view of the web agency, I.
You open up the emails at eight 30 in the morning and there's your list of all the changes that your clients wanna see that day, that typical process means reading the email, typing in the URL of their website, logging into WP admin, finding the piece of content that they're talking about, which honestly they've probably forgotten to tell you about.
And so you have to re-mail them to say, which page was this change you wanted on? And, and then, find it. Publish the change that they've, perfectly encapsulated that they want. Of course. click save, rewrite an email basically saying, okay, I've done it. Please have a check, blah, blah, blah.
Okay? It's not the end of the world. It's not like an insurmountable problem, but write there in, in 30 seconds, I've described five or eight minutes of my life with your solution. You just log in and click, and if it's okay, if the AI suggestion is okay, okay, do you just click publish or something and it's just done.
It goes to the right page, it knows where to go, rewrites it, and then that's it. Does the client get a notification? what happens?
[00:12:06] Zach Hendershot: Yeah, they get a notification, but that's exactly it. Yeah. I mean it's, clicking a single button, right? You, get dropped into the dashboard in the Rooney, plugin. I. The dashboard lists all of your, stories, all the edit requests that came in that require work. You go one by one and you click the button.
As long as everything looks good, it's our mission to make that as efficient and fast as possible. Everything that we've done, and you've described the very simple capture tool, right? There are lots of capture tools out there with 10 different tools for different use cases. We're trying to make this as.
as simple as possible. We know dealing with clients and sometimes less tech, technical savvy clients can be challenging. And so it's really about how do we make this simple, fast, effective, efficient and just get out of your way so that you can just get the work done and make the updates and make it a better experience, for everybody involved.
That's been our mission the whole
[00:12:56] Nathan Wrigley: So is
[00:12:57] Zach Hendershot: and AI is a great tool to help enable that.
[00:12:59] Nathan Wrigley: So is Miruni always capturing an image then, or are there other options? could you, I don't know. the video that I saw it, it was an image, so I just assumed that it would always be an image. But is have I gapped got that right.
[00:13:12] Zach Hendershot: Yeah, so today it's, just an image. the, original tool was a feedback tool you alluded to, this being in market be beforehand and it has been as a traditional capture tool. but for this edit request automation, right now it's a capture tool and that's been very intentional because the screenshot will give us.
The vast majority of the information we need to implement the request. But we do have on our, on our roadmap, video and audio, as more enhanced, more sort of expansive ways to describe the problem, especially for more complicated, maybe interaction changes being requested. Sometimes video is a good tool for that.
but we also wanna also enable the, audio and audio transcription to make it even easier. Maybe you don't want to type out a longer request, you just want to. verbalize it. We can transcribe that and, we'll have that available
[00:14:04] Nathan Wrigley: Okay, so Miruni, as it stands at the moment, we're recording this at the beginning of April, 2025. this is very much a tool where the client knows the change that they wanna make. So it's, I wanna do, I wanna make A into BI. And I know what that is and I know where it is and, here's the screenshot which demonstrates that.
But in the future, you're gonna provide video and audio so that you can have more nuance and more of a conversation there. So rather than swap A for B, it's, would it be better if blue became red or picture of cat became hamster or whatever it may be that there's more of a conversation can develop from those kind of things.
It's not just a binary, swap this to this. Is that kind of the intention there?
[00:14:47] Zach Hendershot: Yeah, exactly. We wanna support one more of a conversation around, the requests being made, but also more, complex requests as well, which may include multiple changes across a page. It may include the creation of new content blocks, for a podcast episode that you want to highlight on your site.
It requires maybe a new page type with, that maintains a lot of the similar page structure and theming, but it's a new structure of content. those require a little bit more sophistication in implementing. And so maybe a video or a, audio might, give us the ability to capture the nuance of the
request So that's very much a big part of it.
[00:15:26] Nathan Wrigley: Wait and see how that comes. But, you can go and check the video. I'll put the, like I said, I'll put the video that I received into the show notes, the link to that, so you'll be able to capture that. I'm just wondering how this works in terms of whether it's SAS or all plugin. So for example, if I'm an agency owner, do I need to install this on a one-to-one basis in every client site that I've got?
Or can I, put that plugin on a. I had a dedicated site, which is for my agency, and then a child plugin can go on the, client sites, things like that. So just talk around how, does the plugin work and do you have a SaaS that it binds to or is it all plugin?
[00:16:04] Zach Hendershot: it does have a saa it binds to, today and it's a, it's a web application, but the way that you enable it right now is a one-to-one relationship. So you install the plugin on the site that you want to enable these automated edit requests on. and those, when those requests come in, those go to the SaaS, platform that we have.
That's at the, word, the WordPress plugin. Pulls the data from that SaaS and then exposes all the data. all that mechanism is, operationalized for lack of better term through, that SaaS product. But very much on our radar is what you alluded to earlier, which is really exciting place to be, which is we wanna orchestrate website updates across a portfolio of websites in a centralized place.
So in the future. in the next few releases, we wanna enable the ability for you to log into one place, app.marina.io, and push edit requests, to a portfolio of sites. so you have the one place to log in and you can manage the sites across your entire portfolio, whether that's, five or 10 'cause you're a freelancer or, a hundred if you're a larger agency.
so that's very much on our radar. We know a part of. making a lot of this work easier and faster is orchestrating multiple site updates across, across a number of sites. And so that's a big part of our, roadmap looking forward.
[00:17:24] Nathan Wrigley: Is the, intention of this product always gonna be content changes? So the, website's already built, it's been shipped, the client has access to it, or does it lean into. I don't know, development of a website, we've started with this kind of wire frame approach or something like that.
And would it be bent or is that kind of really stretching what it's capable of? I.
[00:17:48] Zach Hendershot: It is not stretching what it's capable of. And I think you're hitting on a really exciting place for, where was he the future or Rooney. We do believe, the initial big pain point to solve is content updates, image updates, the evolution of a site that already exists. but I gave you a little bit of a, a teaser of where we wanna go and that is.
Creating new content blocks, creating new pages, maintaining consistency and look and feel, and a lot of the strategy work that has gone into creating the experience that you see and extending it. we don't, we're not really focused right now on net new creation, like standing up a brand new site or doing anything from zero to one.
We're focused on what I've been saying, 80 to a hundred, right? The, long tail of how the site needs to evolve to meet the changing needs of the business or the, product or whatever it is that you're building the site for, which is often measured in years and not the months that the initial creation of the site may, may take.
And so there's lots of things that fall into that creation of new pages, creation of new, content blocks, evolving content that exists on the site today. but it also involves managing. the stakeholders, over time. So communication with stakeholders, managing the requests that come in, obviously from an edit request perspective, but also just like communication and collaboration in general.
so we see the, a broadening of our capabilities in terms of different types of content, in different types of content creation, but we also see like the operating system of an agency is, something that we think. Some of the tools that we're building and some of the capabilities around AI that, that are, only becoming more and more prevalent.
we can help automate so many parts of how you just run your business as an agency or a freelancer. That we want to start to look at different ways we can do that and help make things faster and more efficient and ultimately save a ton of money. we know agencies and freelancers live and die by the hour, right?
And so every hour we can save you is, more money in your pocket, at the end of the day. So that's what we're focused on.
[00:19:58] Nathan Wrigley: the fairly profound thing here is that if we'd have been able to rewind the clock, a decade is a long time, but maybe even four years or something like that, three years maybe. just even the notion that you could get an image. Just present it to some thing online. We're gonna call them l and m's.
Now AI, for want of a better word. And it could just do all these things like pause the text in there because it's not like you need to encapsulate the whole paragraph. Even in the image, you can just take a bit of a paragraph and it could be a messy bit of the paragraph, the client doesn't necessarily need to, okay.
I've gotta get it exactly lined up so I capture the paragraph perfectly. They can just. Just drag it fairly clumsily and the LLM will just read it and then do these suggestions on the inside. just the fact that's possible firstly is a miracle, but then that it can be wrapped up in a myriad different ways by, all the different SaaS products that you can see, but also.
Your innovation, I find to be really curious. I have, a question though. in the example that we've been describing so far, I talked about images and things like that, but we, focus largely on text. So I'm a block editor fan. I use the word, call WordPress blocks often, or a few blocks that come from some third parties.
And often it'll be text headings, maybe some lists, things like that. But it's a fairly uninteresting cornucopia of things that I'm using. So that I guess, is fairly easy for your AI to pause a little bit, but what if I'm, I don't know, what if I'm using elemental. Or what if I'm using some other page builder where things might not be quite the same, they might style their things so that they look like a thing, but they're not a thing.
What I'm basically asking in a very roundabout, long-winded way is, can it do things in the wider WordPress ecosystem outside of just paragraphs, headings, texts, that kind of thing.
[00:22:02] Zach Hendershot: The, exciting answer for us is yes, and this has been a lot of hard work for us to make this, as seamless as possible to accomplish its mission, I've been saying of fast and efficient and easy. yeah, right outta the gate, we support, built-in blocks in Gutenberg and, WordPress, and then we support also.
Elementor being one of the biggest page builders out there. and we can manage all sorts of, wide range of sets of content that are managed within these page builders. in Elementor, for example, an example I really love to share is that elementary, if you're familiar with some of the internals of elementary works, they have a, big JSON file that is basically a big configuration file for the theme that is running on top of page builder for configuring maybe, charts and graphs and, slideshows and other, sort of more advanced, visualization capabilities, we can go in and directly modify individual elements of that configuration file that JSON file in a successful way that implements the request being made. My, my favorite sort of simple example, but I think powerful example is, we have a demo template that has a pie chart and you can, it has three or four different configurable slices of the pie chart that ap add up to a hundred. You can highlight the pie chart. say I want the slice that is currently 14% to read 18%, and the AI will go in, find that deep in the j.
S. Configuration file for Elementor, change that. And then if you don't give it any other information, it will equally remove the percentage point from the other, items. Or you can just explicitly tell it, I want this to be 14, this 16, this 28, whatever it is, and it will update the rendering of that pie chart, exactly how you
[00:23:50] Nathan Wrigley: do you as the agency need to explain? To the LLM in some configuration. That's o in Miruni. Do you need to say this site is an elemental site, so when you are going out? No, it just, looks at the do and figures out. Okay. This is a typical, I dunno, a class name that elemental use as a wrap for their, whatever that thing is, that chart.
[00:24:15] Zach Hendershot: We, can infer, exactly what, what's actively in used, even if you have Elementor and maybe a Beaver Builder or whatever plugins installed. It's not just based simply on the plugins that are installed in your system. it's what the theme is actively using, to render whatever content is that you are, that you captured in the, in the screenshot.
So it will know based on how, where it found that content and what's rendering that content, that this is, generated through Elementor. And it will go, and, In the places it expects to find things through elementary based on that knowledge and make the change being requested. So it, infers all the things it needs to know to make the change as effectively as possible.
[00:24:53] Nathan Wrigley: It's not, it's absolutely not. The, the other question I've got is obviously if I were to go into the exact page that the client has described, client has sent me the request, we wanna change this word to this word, blah, blah, blah. your AI's figured it out. If I was to go into that page and edit it, I would click update.
it's already been published. I'd update that and the entire page would be resent to the server so that it can, it, the, whole thing goes. you don't just send the paragraph on in question. So that's the question. When you are saving something, are you re rendering? I'm presuming you're re rendering the whole thing.
I don't know how that's working. if you're getting into the database and figuring it on that level, or if you somehow. I don't know, using the rest API or something like that to save things. Tell me, I dunno.
[00:25:44] Zach Hendershot: I, think you're asking about, how we generate the preview. Is that
what of the
[00:25:48] Nathan Wrigley: I'm asking that if I am just wanting to change one word in one paragraph, the UI only shows me the paragraph in question. It shows me what it was, what it will become, and it feels, at that point we are only updating the paragraph. But obviously in the WordPress world, we don't just update the paragraph, we, update the whole content.
So I'm presuming you are doing the same thing as well. Sorry, that was badly explained.
[00:26:12] Zach Hendershot: Yeah, no, we are exactly doing that. Like we are, we're figuring out where that content is making the update directly, and then we do, when we generate the preview, we know the page that the request came in on, so we render the page. So you can see it in context, but you can also click around on that preview and see.
I. The request to change how it was applied to different parts of the site, particularly like menu items or other requests that may have like cross site impacts. we generate a draft that allows you to see that impact across a number of different pages and, do it like a human would do it just in an automated way, right?
So all the things that you understand and know how it works, doing it manually, we're also doing it quote unquote, manually. just facilitated and enabled by, an AI agent.
[00:26:56] Nathan Wrigley: Yeah, that's, so that's an interesting point. And I think, that's how I like my, ai. I think I like my all AI with a bit of human in it, an actual human, not like humanity faked by an ai. And that is something that I really like about this tool. The AI is the suggestion. It's, it basically, it's doing the suggestion it, look, the client said this.
I think this, I, as in the ai, we think this, is what we predict is, the change that's required, but it cannot commit it. The human at the agency has to go, yep, that's the one or no. And that I think is, very intelligent. 'cause I think handing this over directly to AI, maybe.
[00:27:39] Zach Hendershot: Yeah. and I want to comment on, something that you said a little bit earlier in that question as well. Around human in the loop, is what we call it, we're committed to human in the loop. my, my belief very much is that AI is. Is an enabler of, human, thought and human action, right? And not a replacement of it.
And so a lot of the strategies that you're seeing in Rooney really is to make us better, faster, and more efficient and more effective, not to pull us outta the loop, right? Not to like. Automate everything automatically for us with no sort of say in the middle of it. so our, workflow is very much in support of that and how we manage when it gets it wrong is a good kind of targeted example of that.
to answer your question directly. The AI will not always get it wrong, will not always get it right, I should say let's see. They're already influencing the words that I'm saying. they, they don't always get it right. And, you know this from using Chatbots Chat, GPT and Claude or whatever, it doesn't always get the precision or understand the context perfectly. So we've helped that problem with a couple different tools and more tools to come. The immediate tools we have available is a retry. We've actually found. Just clicking a retry button, just do it again.
Basically. actually redu, results in significantly better, resolution to the issue, in one to two retries at the most. so we get a large, we close a large gap just by allowing the AI to retry. and it's a retry in the sense of you got it wrong. Try again with the same context you had.
So just knowing the lm knowing that its initial, recommendation was incorrect, forces it to deviate from the, from the recommendation so that solves a big problem. Very simple. One button click. we actually, I. Internally have validation phases of the AI that will validate whether it generated valid JSON or a valid file or a page that didn't break it.
that we don't even expose the AI suggestion to the end user if we, if it failed those validations, and we'll have it retry, up to three times to get it right. we can't always detect every problem internally to our system. And so if it does get to the end user that it's incorrect or just is semantically wrong or spel, the brand name is spelled differently or incorrectly, things that it would be hard, mechanically for us to detect.
we allow the retry button and then what we'll have in the roadmap. Is, very soon actually is a chat interface. So the website admin can say, can give, English level or, language level descriptions of what it once tweaked. Maybe we, Gave it copy that was too flamboyant or too over the top in terms of the description of the, of a product that you're launching or whatever.
And say, you can basically say, tone it down and try again. So try again wouldn't solve the problem because it would just do the same thing in a different format. But you can tell it, tone down the language of the, use less marketing language here and it will retry with that additional context and, try again.
Or if it's broken, We missed. Some page breaking activity or whatever you can say, Hey, like this page is broken in this way. Retry the edit and, have a conversation with, the request. And, hopefully, in large percentage of cases, ultimately get it right while still saving you all the time it would take to go through all the pages and posts and find it and update it and edited and everything else manually.
[00:31:12] Nathan Wrigley: Can I just ask then, in your experience, so in the length of time that this product has been alive, let's just put our finger in the air a little bit. And figure out how much time typically might be saved. obviously, the task at hand will matter enormously, but what kind of numbers are you throwing out there in your marketing as this is the, kind of order of magnitude that we're trying to save you in terms of time here.
[00:31:39] Zach Hendershot: Yeah, we, we talked to a lot of agencies in building this product in early phases, and we did some data collection measurement. So I actually feel somewhat confident in the response I'm about to give you. But what we say in the marketing perspective is about, we say it'd be about 80%, of the hours it would take, the time it would take to go and make manual updates for all the requests that come in.
And that's, that's a across a wide range of different types of edits. To be specific, the content, like the word change, right? Like a misspelling or a content edit,
[00:32:10] Nathan Wrigley: That's
brutal, isn't it? Yeah. one
letter.
[00:32:14] Zach Hendershot: the interesting thing about that is while our approach, it may take a. 30 seconds to just approve, look at it briefly in the preview, click approve and it's live on your site.
That's maybe less than 30 seconds. Doing it manually may take you a minute and a half, couple minutes maybe in the, best case. Often those are higher volume, right? So when you're adding up 30 seconds versus a few minutes over 10 a week or 20 a week or 30 a week, right? That starts to become meaningful, meaningful numbers.
But I think the real magic and that's great and we love that and we support that use case. And that's a very fast use case. There is a, another set of use cases, which is. More complicated swapping images or, adding content blocks or, changing hours. You deep inside of a, more complex page structure, where you don't even know where it is and you have to go and find it 'cause you don't update often.
which is what we heard over and over again. I don't do. There's a, wide range of types of updates, and I don't do all of them frequently enough to know precisely where to go into the WordPress site to change it. And so I spend often 10, 20, 30 minutes, if not a few hours on these what might seem like relatively simple content updates that we can automate again in less than 30 seconds.
where you may be spending on, on. an hour or two hours, whatever it is, doing some of these updates, mostly hunting and pecking. and then when you start talking about structuring pages and content and other things that are aligned with the theme, we can automate a huge chunk of that.
But that's a lot of, I. Thought energy that goes into it that also may take a few hours to go and do that is again, 30 seconds a minute, using the Miruni process. So, when you start adding that up over tens and maybe hundreds for larger agencies of requests in a week or a month, that starts to be hundreds if not thousands of hours, over a period of time that, were, that we're able to save.
[00:34:19] Nathan Wrigley: If the person listening to this is feeling that pain, then they are your ideal, So if, it's sending things to an ai, then presumably there's a cost to be born for that, and that cost obviously needs to be passed on by you. let's just quickly get onto the pricing. So caveat mTOR, the pricing is.
Maybe it'll change, who knows? But at the moment, we've got a freelancer, edition, which can I just commend you on ending a price in a zero because it's,
[00:34:50] Zach Hendershot: That was an intentional
[00:34:51] Nathan Wrigley: a nine or a seven. nobody's getting confused. By 49, it's $50. So anyway, it's $50. And the constraints on that, and this is where I want to go.
a hundred SMART edits, 25, active projects. Unlimited request captures unlimited members and guests. Let's go back to the smart edits. what are these hundred smart edits? What counts as one of those?
[00:35:16] Zach Hendershot: So smart edits are the automated, AI a edits that we've been talking about. So that's our marketing name for them. Smart edits. our pricing is very simple and I've been in the SaaS space for a long time. I've been buyers of SaaS products, and so I wanted to call out a couple things that we've been very intentional about.
One is. transparent pricing. A part of that is not having the tricks of, 49.99 a month or whatever. So that's, a thing. but the other is very simple, consumable based pricing. And the only real variable between our tiers, which start at $50 a month, is the number of smart edits you get.
And a very small, restriction around the number of active projects. those are the only two variables right now that exist, in our, pricing structure. So you're basically paying for more smart edits and you consume a smart edit when you actively apply the suggested change. When you click the apply button, that's how you consume it.
You can continue. You can continue to get the recommendations and it won't save you a ton of time if you're not applying it. But I think it is maybe still helpful in some ways to see how we might make a recommendation to make a change. and we do provide some tools to see the code Diff that we're making and some of the backend stuff that we're, doing of it.
But it's really about consuming the
smart edit.
[00:36:33] Nathan Wrigley: so let me just pause that again. if a client is trigger happy and just sends off like a, bazillion, little snippets, I wanna change this, I wanna change this, I wanna change this, and 50 50, suggestions later, they're all in the backend. I'm not paying for those until I commit them.
One com a commit is a smart edit.
[00:36:55] Zach Hendershot: Yep. That's when you click the apply button. So yeah, the, you can get a hundred from an individual,
[00:37:01] Nathan Wrigley: And I will. Yeah.
[00:37:03] Zach Hendershot: which you will. Yeah, exactly. and you can make a decision around, one, whether you want to consume a smart edit for that, or if you don't.
we want to be, again, an enhancer and enabler, not a replacement for you. So that's, again, giving you the power of making the
[00:37:18] Nathan Wrigley: Okay, so you don't need to rate, limit your clients. That was really what I was worried about. If the 'cause, because what's interesting there is you are still taking a hit on that as a company. You are sending that to the LLM, getting the feedback from the LLM, putting it into the ui. So that's interesting.
I like the that you've got that way round because I could telling my clients, no, you can only click that button four more times this month. No more
[00:37:43] Zach Hendershot: That was the conversation. Yeah. That was the conversation we didn't want you to have. and that does mean that, we had to, model our pricing in a way that allows for that so that you can have unrestricted, client access to this tool without, without, putting the burden on.
The customer at the end of the day. Again, this is all about providing a great experience for your end user and customer upleveling, the experience you can give them. and that's not a conversation we wanted you to have. And we've done the cost modeling on our side to make sure that we, and we can cover that for now.
[00:38:17] Nathan Wrigley: freelancer was $50 for a hundred Smart Edits Pro, the pricing, it just says contact us, but you, there are other tiers, 502,000 and so on. Do I have to flip flop between those tiers or can I do a la carte if I go for 101, can I buy like 10 more or 50 more or
[00:38:36] Zach Hendershot: One, once you're at a tier. Yeah. We're gonna offer, packages of smart edits, above and beyond your tier. the next tier. the, business tier is more made for like medium, size agencies who, might want to have or will require more smart edits. They can also buy packages as well.
Right now we're in a mode of creating tailored packages for their workload. We found a lot of variability, in terms of, medium agencies, medium sized agencies that have. 10 very large, highly active projects, and then some that have a thousand really small, low volume projects. and so we're right now trying to figure out how to meet everybody's needs and the best we can.
We know freelancers, and that, that initial, plan tier is, is pretty well understood. We're working individually with, agencies to get them the right plan that's both cost effective, but, but gets them the support that they need.
[00:39:33] Nathan Wrigley: very nice. Honestly, what an interesting product. I always a bit fearful of ai, but I love the balance here of AI mixed with humans and, saving actual humans time at the same time. yeah, profound. Congratulations. the URL is Mai. uni. M-I-R-U-N i.io. I hope I got that right.
It is io, isn't it? Yeah. so dear listener, go and check that out. Zach, it only remains for me to say best of luck. I hope it's, an unmitigated success and congratulations for what you've done so far. Thank you very much for chatting to me today.
[00:40:10] Zach Hendershot: Thank you for chatting with me. I love talking about what we're doing, so I appreciate it.
[00:40:13] Nathan Wrigley: No worries.
Okay, that's all I've got for you this week. Hopefully you enjoyed that. If you're interested in Miruni and what Zach had to say, and you've got some comments on that, please head to wpbuilds.com and search for episode number 422, and leave us a comment there. We'd really appreciate it.
The WP Builds podcast is brought to you today by GoDaddy Pro. GoDaddy Pro, the home of managed WordPress hosting that includes free domain, SSL and 24 7 support. Bundle that with The Hub by GoDaddy Pro to unlock more free benefits to manage multiple sites in one place, invoice clients and get 30% off new purchases. Find out more at go.me/wpbuilds.
The lights are kept on this week also by Bluehost. Bluehost, redefine your web hosting experience with Bluehost Cloud. Managed WordPress hosting that comes with lightning fast websites, 100% network uptime, and 24 7 priority support. With Bluehost Cloud the possibilities out of this world, experience it today at bluehost.com/cloud.
And Omnisend are helping us with our podcast this week, Omnisend, do you sell your stuff online? Then meet Omnisend. Yes, that Omnisend. The email and SMS tool that helps you make 73 bucks for every dollar spent. The one that's so good, it's almost boring. Hate the excitement of rollercoaster sales? Prefer a steady line going up? Try Omnisend today at omnissend.com.
And sincere thanks, go to GoDaddy Pro. Bluehost and Omnisend for their support of the WP Builds podcast.
Okay, we're about to close out this episode, but just before we do, don't forget if you want to advertise on WP Builds like the three fine companies you just heard from, head to wpbuilds.com/advertise to find out more.
Okay, I am gonna fade in some cheesy music generated by AI, which is very appropriate for this episode. I'm gonna say stay safe. Have a good week. Bye-bye for now.
