439 – AI and automation in website audits: Pradeep Sonawane talks about Web Auditor

[00:00:20] Nathan Wrigley: Hello there, and welcome once again to the WP Builds podcast. You've reached episode number 439, entitled AI automation in website audits, Pradeep Sonawane talks about Web Auditor. It was published on Thursday, the 2nd of October, 2025.

My name's Nathan Wrigley, and before we get to the interview all about Web Auditor, a few little bits of housekeeping if you like what we do at WP Builds, why not subscribe, keep updated with all that we do, head to wpbuilds.com/subscribe. Fill out the boxes, follow us on social networks. You know the drill, do all of those bits and pieces, and we will keep you updated when two bits of content emerge.

The first one is every Thursday we will produce a podcast episode. That is what you are listening to now. Typically it's an interview with me and one other person. And then on a Monday we record the This Week in WordPress show, which then gets recycled as a podcast episode, and we'll notify you on Tuesday about that. That's really all that we'll do, and if you subscribe to our newsletter, fill in your email on the website there, and we will keep you updated about that.

The other thing to mention is that Black Friday is coming around pretty soon. I will soon have our Black Friday page up and running. So if you are a WordPresser and you have a plugin, a theme, a block, anything in the WordPress space, which you would like to sell into the market over Black Friday, in the very near future, head to wpbuilds.com/black, and we'll have opportunities there for you to let us know about your product, well in advance.

We have a page which has absolutely tons of traffic in the run up to Black Friday, and so hopefully we'll have that page up and running pretty soon. If you go over there, you'll be able to find an Add Your Deal, or Add A Deal button. Click on that, and yes, you will be able to add your deal so that you can be discovered in what is a very, very busy, tumultuous period.

Also, if you are listening to this podcast, you may well notice that we slot adverts in from time to time. Those are our sponsor slots. The reason that we have those is to keep the podcast going, but also to enable the sponsors to have their messages in front of you, a WordPress audience member. If, like I said a moment ago about Black Friday, if you have a product or service you would like to have in front of the podcast audience, head to wpbuilds.com/advertise to find out more.

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 a lot 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.

We're also helped out this week 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 are outta this world. Experience it today at bluehost.com/cloud.

And we're also joined by Omnisend. 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? Well try Omnisend today at omnisend.com.

And dear listener, we would like to thank GoDaddy Pro, Bluehost and Omnisend for their support of the WP Builds podcast. Podcasts like this cannot happen without the gracious efforts of companies like them.

If you would like, as they did, to get your message in front of our audience, head to wpbuilds.com/advertised to find out more.

Okay. What have we got for you today? Well, today I'm talking to Pradeep, all about Web Auditor.

This is a brand new tool. It's come out of rtCamp, which is an Indian enterprise agency.

We talk about what the tool does, how it all started. It started as a CLI tool.

We talk about the state of WordPress in India, and how it's going gangbusters. And we really dig into the tool, what it can do. What it enables you to do for your clients. What you can put on autopilot. And what are the plans for the future.

If you've got problems with websites, and you would like a very automated way to figure out what the heck is going on, and why things are breaking, this episode certainly is for you.

I hope that you enjoy it.

I am joined on the podcast by Pradeep Sonawane. Hello Pradeep.

[00:05:16] Pradeep Sonawane: Hello.

[00:05:16] Nathan Wrigley: We've had a little bit of back and forth before this because I'm, really unable to, say your name and I apologize for butchering that at the start, but, you've been very gracious allowing me to say it in the way that I have said it. You are joining me today and we're going to talk about something called, web Auditor.

And can I just, ask the listener, why don't you just pause this podcast right away before we get stuck into too much and go and check it out. You're going to find it at webauditor.io. So that's W E B A U D I T O R dot io. That'll probably make this podcast a whole lot easier. Because you'll be able to have had a poke around and understand what the tool is for and how it works and what have you.

So hopefully you've clicked pause and come back and here you are. So let's just get into a little bit about you, Pradeep. Before we begin this, conversation properly, are you able to just tell us a little bit about you, perhaps where you work, and how long you've been in the WordPress space, things like that.

[00:06:17] Pradeep Sonawane: Yes. I just would like to introduce myself. I'm Pradeep Sonawane. I've been working with rtCamp for around, I think 15 years now. I started as a developer in 2009. Over the years I've taken on diverse roles, from project management and pre-sales to project consultancy, and now leading and, building engineering teams at rtCamp.

I have also had an opportunity to work on several Google projects, and I was a speaker at Google's Amp Road Show in 2018. It happened in Mumbai, and it is a tech behind web, stories if somebody doesn't know about it. Most recently I've created web auditor io. It's a SaaS tool that helps developers and website owners identify the performance, bottlenecks and experiment with the solutions. Making the website optimization faster and most, more insights, insightful. So yeah, that's a little bit about me.

[00:07:23] Nathan Wrigley: Thank you. I appreciate that. I actually, did a little while ago on the Tavern podcast, so I do the Tavern podcast as well. I had, your, I'm going to say CEO, I don't know if that's the case, but I had Rahul, on the podcast and and we got into the whole rtCamp journey and all the different bits, and pieces that they're doing, and what I took away from that conversation, is that WordPress, if WordPress was to be described as alive well and anywhere on the planet, it would be where you live. Your part of the world seems to be really getting into WordPress, not just the code, but the events and the community, and the meetups and we've got the, we've got the whole WordCamp Asia come in next year, which is going to be happening in Mumbai, India.

I don't really know if you want to get into that, but it was just curious to me that the whole thing since COVID, getting events in the WordPress space going again has been really hard, but it seems like over where you live, it's just thriving, which is great.

[00:08:26] Pradeep Sonawane: Yeah, here people would like mostly connect in these events and have a discussion because after COVID it was hard, meeting people and doing networking stuff. So even my team and me, I used to meet once a year, because I stay far from office, and even before COVID we were remote company.

So I used to go, when there is any important event happening, or anybody to meet, or if there is an important meeting, basically. So we think it as an opportunity to meet people. WordPress and WordPress events are very big. Now it's gotten even bigger. So you have many people, experts talking, and you also do networking. You meet, you build your network basically.

[00:09:18] Nathan Wrigley: Yeah, it's amazing. And again, I'll put the link in the show notes to that episode if you want to hear what Rahul said about WordPress and rtCamp and the WordPress community in your part of the world.

But the agenda today is this web auditor tool. Now, I'm curious, I know that rtCamp, I'm just going to summarize it as this, I'm going to summarize it as an enterprise WordPress agency. You are dealing with some. Very large clients, the likes of whom probably many of us have heard before. So you are working at a real scale. Let's just put it that way. You know, where squeezing out every single possible performance gain that you can from a website is probably really crucial.

You've got the kind of clients who don't mess about, they want to know that their website. is the fastest possible thing that it can be. And obviously Googling that, and using Lighthouse and things like that is very important in that.

However, how did this tool come about? How were you afforded the time and the space and the finance to build it? It's not the kind of tool that most agencies have got in their back pocket. What was the itch that you were scratching? And was this something that you did on company time or was it just something that you did as a pet project?

[00:10:32] Pradeep Sonawane: Yes, you are right. I did it as a pet project actually. This was not supposed to be a SaaS tool. It started initially as a CLI tool. This was basically an idea back in, I think, 2013, when I had an optimization project to work on. It was about fixing INP issues, like most Core Web Vitals.

INP was the new metric, that Google added that, in 2013. So I think 2013. So at the time while debugging the website and finding out the issues, the debugging time was too much, and finding the JavaScripts, CSS, and the resources that are being loaded on the site.

And a normal developer would just go inspect the network tab, in the dev tools of Chrome and then find out the JavaScripts or CSS resources, whichever. No, he thinks it's not doing well or taking time. So basically I used to use Chrome's dev tools and block the resources to find out which is impacting the most, like which resource.

Since it took like around one or two hours every time I thought of just making it as an automation or something like CLI tool, because every time I have to perform the same steps, right? So what I did was I started it as a CLI tool that works locally. and then I, basically, this tool was built on top of Lighthouse.

So Lighthouse also gives a CLI to run the websites audit locally. But then, what I was missing in that was to modify the site. I wanted to remove some JavaScripts. I wanted to remove some CSS, and see the impact of that on the performance. And I was not able to get that.

So it was like a manual process every time I have to do. So what I did was I created this tool on top of Lighthouse, used Puppeteer, as a backend, background processing tech. And then I basically asked my team to basically have some custom experiments, some prebuilt experiments.

So something like removing JavaScript is one experiment. Removing render blocking resources is one experiment. Removing low code coverage. So there were three to four experiments like these, and then I used to basically run it and, just like the other tools it runs the audit, but it also runs the experiments and it also runs the Lighthouse on those experiments. And then it compares the experiments with the baseline, to show you like what performance skills you can have with those experiments.

[00:13:21] Nathan Wrigley: That's really curious. So I'm sure that anybody who has developed sites at the kind of level that you are dealing with, but also below that as well, the idea of a web auditing tool where you, I don't know, it might be a SaaS app or something like that. You go and you, plug in the URL of the website that you've just been working on. You press a button and gives you a report, and it tells you useful data about where things are going wrong. Let's say that.

You've gone a step further though, and so you've done the experiments, but then you modify the experiment by let's say, toggling things on and off. So in this case, it might be, okay, let's just switch off JavaScript and see how that flies.

And then you rerun the experiment again and see, it's like an AB test almost. Okay, how did that fix anything? Did that fix anything? If not, okay, we've learned something there. I don't know if anybody else is doing that. I, don't know if you can speak to that. Is that unique, do you think?

[00:14:15] Pradeep Sonawane: So at the time we built, I also didn't think anybody built that. Maybe somebody is doing that right now. I don't know. But, at the time I was building this tool. There was none. Unique in that way.

[00:14:28] Nathan Wrigley: Yeah, really, unique. Okay, let's then go through, because you mentioned that the origins of this was WPCLI, so it's a command line tool, it's text. That's not where we're at now though, because if you go to web auditor.io, like I said at the beginning, and you don't even need to get off the homepage, you'll see on the homepage a video which will run you through the basics of how the tool performs, and the manner in which it presents the data.

But you'll also, see further down information about the different kind of things that you can do and what it's measuring and what have you. And you'll see right away that it's a, it's anything but a command line tool. There's a GUI, you can interact with it, and it basically, it looks beautiful.

How long has it been going from the WPCLI tool that pretty much you could understand and not many other people could, to this wonderful thing that you've built now? How long did that all take?

[00:15:19] Pradeep Sonawane: So it was just a normal CLI, not even WPCLI.

[00:15:24] Nathan Wrigley: Okay.

[00:15:24] Pradeep Sonawane: So it's like how Lighthouse gives you Lighthouse command to run an audit. We had our own command after building this tool, because it runs some experiments also apart from normal audit. Initially it was just a CLI, but when I showed this as a pet project, I was doing in the, in my free time, the team, internal team actually saw this and they liked it. And my goal was to basically just use it internally initially. Just what I did for my project. I wished other team members also use it, like our sales team in our pre-sales activity, or project managers or any team basically.

They can just use, make up this tool and then, see what's, what are the issues. And they can communicate with the clients proactively basically. If they found any issue they can just communicate with the team and then have those fixed earlier.

[00:16:27] Nathan Wrigley: Can I just interrupt then? Just so that I'm getting this story correct. The idea was that, okay, you built it for yourself. You were scratching your own itch, then you thought, showed it to a few team members and they said, actually, this is great. But your intention at that point was for it to be an internal tool for people who you knew were already really credible, and could understand the information, and understand what they were doing with it. Okay. Okay. So that's where we're at. It's an internal tool, probably still quite complicated, but, that definitely isn't where the journey ended. Sorry to interrupt, but carry on.

[00:16:58] Pradeep Sonawane: Yes. So the team liked it and I also showed it to our CEO Rahul. He also liked it. Then he said, let's build it a SaaS tool. So like we can have it, like people can use it. Like wider community can access, the auditing tool like this. Then I said, okay, let's do it.

We did it in I think two, three months after that. And then we launched the first version of this and then we basically got some feedback also. I'm very grateful about that also, because, a lot of core team members, workplace core team members, they actually, responded with the feedback.

And we have addressed those feedback. And then, I also made sure that I tell them that I fixed those issues and then have a look at it again. So they looked at it again. Then they also liked it. I would say it's the feedback that I received and then we address those. And, the improvements that you see in the tool right now, it's because of them.

[00:18:01] Nathan Wrigley: It's there, it's available. It's a SaaS app. Now, as you say, if you go to web auditor.io, you'll be able to see that. Who is the target people for this? What level of understanding do you need to bring in order to make any use of this?

Let's say for example, I'm an experienced developer like you, obviously we know people like you can make sense of this data that's a given. But if we go down the scale, a little bit if you like to, I don't know, somebody who's just implementing WordPress with a page builder. Has no real interest in this kind of field. Does it present information in a way which would be useful to them? Or is there a sort of learning curve where you have to be more developer focused in order to make sense of it?

Does that question make sense? I'm trying to figure out where you are pitching this, at what level of expertise.

[00:18:47] Pradeep Sonawane: Initially I actually wanted to make it for developers, front end developers, the QA, maybe, they can see the performance insights in a very simple, human readable, language. See the report. It was my goal earlier, but then I thought of having the audience more diverse, instead of just focusing on developers.

So then I thought of digital marketers, small business owners, the web agencies, project managers. So I think, later we used AI also so that we can basically have the output, or the language that the report gives. Like initially it was just a, comparison table with some diagnostics issues and that developer would understand.

But digital marketers, the project managers, it would be hard for them to understand. I basically made use of AI, Gemini to be specific. It gives a very nice output actually in a human readable form. People can understand it. Yeah, so the target audience it grew to other people also, like digital marketers and project managers.

[00:19:57] Nathan Wrigley: Okay. So basically what you're saying is that more or less anybody, if you're willing to give it a go. That is a very curious implementation of AI, by the way. I think that's really fascinating. So you're taking, how to describe this? It could be a spreadsheet with numbers on it, let's put it that way.

It could just be a bunch of data, a bunch of charts. That kind of appeals to a certain class of people. They just love all that kinda stuff, and they dig into it and really enjoy it. But then if you're a digital marketer, and you wouldn't know whether that particular cell containing that number was a good thing or a bad thing, if you can use AI to churn out, just text basically, which kind of summarizes it. Then that seems like a really credible use of AI, and also it makes the audience immediately much bigger. Coincidentally, I was playing with some AI the other day and not to do with this particular task, but I was trying to get it to draw out meaningful data from a corpus of information, which I fed it, and I was really surprised at how good it was.

In other words, it could see insights buried in the data, which I would never have anticipated it would've spotted. The data was about a plugin and about the UI and the UX. It could see from the code that certain things would be possible, and I was a bit blown away. So I, guess what, you're saying is the same thing here, it can draw conclusions in a heartbeat that you yourself may not see, which is fascinating. Do you want to speak to that at all?

[00:21:32] Pradeep Sonawane: That's the power of AI, right? You basically need a lot of time to absorb things, right? So if I just give you, if I just show you a table that a developer would understand, I think, but the normal public, normal people, non techy people, I think they won't understand. You have to explain what are these numbers, right?

And how to read that table, the numbers. The summary of that table was actually the context of my AI prompt. Basically I passed that as a context, and then AI basically made sense of that data and then it gave the response, which anybody can understand.

So one example of that could be it can say that your JavaScript experiment is performing better than your baseline because it has, so many scripts and it's basically potential savings of, let's say, two MB of JS files, so something like. So this language anybody can understand, but just looking at the numbers, you won't understand.

[00:22:36] Nathan Wrigley: Yeah, that's interesting and also what I've noticed, it's very good at summarizing a corpus of information. So let's say that you've got, I don't know, in your case, a spreadsheet with eight columns and 55 rows. You would have to individually go through each one of those, and make a judgment about each one of them.

And, obviously the AI on some level is doing that same task, but it does it in a heartbeat. And what I found it's really good at is not talking about the stuff that doesn't matter, and focusing in on the bit, the one bit, the two bits, the three bits, whatever it may be, where there's enough change that it's significant to talk about. And, so it just hones right in on the bit that, you need to focus on.

Okay. That's really, interesting, and a fascinating use. Kind of the perfect use of AI. I've got this feeling that generative AI, where it's trying to slurp stuff in from the whole internet seems to just hallucinate most of the time.

But in this case, you are just giving it a very finite, walled amount of data, a small corpus of information, and it's working on that. And then comparing it to the results from 10 minutes ago. And at that I think AI is really good.

[00:23:49] Pradeep Sonawane: The AI is basically, is going to respond with the output, which would be as good as your prompts, right? If you have good prompts, it's going to be better at the response. So I made sure that, okay, we pass the context well. The minimized version of the data. Let's say I just wanted to compare the performance between the baseline and the other experiments. I will just pass the numbers. If I need the detailed analysis of JavaScripts used in the website. And then what I did with the experiment version of that is I removed the JavaScript and then see what exactly is benefiting me in the experiment, and what is the impact of that.

Such things actually, if we pass that information well, as part of the context, it can understand it and then it can respond better.

[00:24:49] Nathan Wrigley: So it's got this context of that was what it was, this is what it is now.

[00:24:52] Pradeep Sonawane: So this is for just comparison. Now I have taken it further, in terms of analyzing the issues. The lighthouse actually, it has its own analysis, right? It can you priorities, like top priority issues in the diagnostics. And then, there are low priority issues. So basically we pass the top priority issues also with the context with the actual elements, let's say issues. In the LCP issues you get LCP element. That element also we pass to AI and then responds with the exact solution that you'll need to follow and fix that issue. So it's curated, you can say tailored response just for your issues, your website issues.

[00:25:42] Nathan Wrigley: Okay, great. So let's just rewind a little bit and imagine that I'm a first time user. I've come to, webauditor.io. And I'm logged in and what have you, and we can worry about how you set up an accountant, pricing and things like that in a minute. But for now, if I go in, how does it work?

So obviously rtCamp, it has a heritage of working with WordPress. This is a WordPress podcast. Is there like any kind of companion between this tool and WordPress? In other words, is there like a plugin which binds to it, or is it a complete SaaS solution? And does it do anything WordPress specific?

Is your auditing tool going out? Okay, we found out this is a WordPress site. Let's try this laundry list of typical WordPress things that often go wrong. What I'm asking basically is does this have any secret sauce when it comes to WordPress and it.

[00:26:33] Pradeep Sonawane: So what this tool does is basically it finds out the technology used in the website. So let's say your website is using WordPress, any CDN services, any plugins basically.

We also, basically have that, when we pass it to the AI, as a prompt or as an input. We pass this information also so that we get the solution based on this.

So it's like AI will respond in terms of WordPress specific sites. Look, if I am auditing. WordPress site, then it'll respond with the solutions that are more appropriate for the WordPress. And right now it's a SaaS tool. But I'm also thinking of, we are thinking of building a plugin that connects the SaaS tool with WordPress, so that people can have, in their backend, they can basically audit the site. They can have all the features from our SaaS tool that we have, like recurring audits and other things.

They can have the dashboard in the WordPress itself. So we have our own dashboard for our SaaS tool. So basically it shows how many sites we audited in this month and what is the performance, et cetera. So that we can see in WordPress.

It's my, actually wish to have that plugin. So we'll soon be starting working on that. And, so let's see.

[00:28:09] Nathan Wrigley: The name says it all. When I think of the word audit, I'm not thinking about we are going to make changes. I'm thinking about, we're going to describe what could be changed. That's what an audit means to me. It's, I'm an expert. Here's all the things that I've discovered about your thing. If, I don't know, if I audit a Word document, I'm not making changes. I'm looking at it and making suggestions, and that feels like what this tool is.

However, interestingly, having a companion, let's say WordPress plugin, it would be nice, and maybe this is what you're describing, it would be nice if you could implement the suggested modification that web auditor has. If you could, I don't know, if you were confident in the tool and you could just click a button, and it would then perform the action to make the improvement on your behalf. That would be crazy useful, wouldn't it? That would nice.

[00:29:02] Pradeep Sonawane: Yes. Yes. And yeah, I think that's actually I would say that again, power of WordPress plugin also, because when we audit a normal site on our SaaS tool, it's nothing. We just read, analysis. We can see recommendations and all. But developer will still have to work on those fixes, right?

But with WordPress plugin, we can identify the issues with the plugins also, and then we can have a proper fix. Or we can say prebuilt, predefined fixes for such issues. So then let's say 90% of the people are having the same issue, that can be fixed via just a plugin that we might have.

And then, this basically WordPress plugin gives us power to add some extra code, extra hooks and filters that we can just plug in. And then ask basically plugin itself to fix those issues.

[00:30:03] Nathan Wrigley: Then you could do a whole other round and just check that the applied solution has actually done what it thought it was going to do. Yeah. Yeah. That's really interesting. Really interesting. So maybe that's more of a roadmap, but certainly possible and a really intriguing thing.

Okay, let's go back to where we're at the moment. So again, I've logged in. Presumably there's a field which somewhere says, what's the URL of the site that you want to get started with? I'll copy and paste the URL of the website in question.

From that moment, what are the things that you are currently, so we're recording this in September, 2025, obviously subject to change and roadmaps and things like that. But at the moment, what are the key things that you are analyzing and giving audit information on? There might be hundreds or dozens of those things, but just the top tier items that most people would be interested in hearing about.

[00:30:53] Pradeep Sonawane: Basically when you enter the URL, it starts auditing. It basically checks the baseline Lighthouse performance report. So that you would normally get in your Chrome browser also.

[00:31:06] Nathan Wrigley: Yep.

[00:31:06] Pradeep Sonawane: And by default we are doing these experiments for each audit. So it's not like it's an optional thing, or user will have to click somewhere else also, do extra things for that.

On just one button click, you get the baseline audit. You get the experiments also run, and those experiments also are passed to the Lighthouse audit. And then it can compare the experiments with the baseline. So everything happens in just one button click, under one minute, mostly in one minute.

[00:31:42] Nathan Wrigley: And the experiments that you are talking about there, this is you trying different things out, and then rerunning the audit with those modifications made. Could you just describe what are some of the different bits and pieces that you are bringing in with your experience? I imagine it's, I don't know, switching off certain JavaScript components of the website, this kind of thing. So just run through some of those so we get an idea.

[00:32:06] Pradeep Sonawane: Yeah, so let's talk about just disabling JavaScript, right? A normal person will just go and disable or block all the JavaScript in the network panel, and then run the lighthouse again. However, that's a simple experiment, right? Nobody would, in the real world, nobody will do that.

But it's just to find out, what impact it has on my website. I would do that. It's certainly not affecting the real website. It's just happening in the Chrome instance that we run in the controlled environment, and it's not affecting the real site, right? What we do is we just block the JavaScripts. All the JS patterns, and then it basically has a website with images, maybe videos, and CSS, no JavaScript. With that we just run the Lighthouse. It automatically runs it, and then it checks the performance of that. Naturally it should be better than the baseline because you are removing.

[00:33:07] Nathan Wrigley: Yeah.

[00:33:09] Pradeep Sonawane: Yeah. If you are removing all the JavaScript, obviously it should be better than your baseline. Like it should be like that only because, in my experience, over the years, I found that the JavaScript is like top most issue in the performance that we see. When you don't have JavaScript, you get better scores, basically.

But then in real world it doesn't happen. You have to use JavaScript. You have to keep those. Let's say we just want to have first party JavaScript. Means our own site's javaScript, not the third party. So we have a experiment which removes third party JavaScript, and it keeps first party JavaScript. So your website JavaScripts are there, perfectly running fine. And then you can see the performance and you can compare it with the no JavaScript experiment, and with the baseline also.

So that gives more insights, basically like how your first party JavaScripts are performing. How your site is performing without JavaScript. Then we have render blocking resources also. So we get a list of render blocking resources. CSS is basically render blocking. Then some JavaScripts would be render blocking if you don't use a sync or defer attribute while loading them.

Those JavaScript or those resources, we automatically disable and then the Lighthouse is also run on that. Then we compare that with the earlier experiments and with the baseline also.

So it's like progressively we are doing the experiments and we are comparing individually with each of them. Basically it's like how developer would do it in the Chrome instance, in his Chrome, in the dev tools. The same way we do it. But with Puppeteer instance, it's a headless that it provides, right? You can use Puppeteer and then headless chrome can be used in that. You don't see the actual things there, like what is happening, but, you write script for that. So that's happening in the background.

[00:35:16] Nathan Wrigley: So you basically click go basically, and, then it'll do things in the background and give you the results on the other end.

[00:35:22] Pradeep Sonawane: Yes, and one more thing is, after this, like you would say, okay, I don't want to disable all the JavaScript or all the third party. I just want to disable a specific JavaScript that, like it's not usual JavaScript you would say. So you get an option basically in our tool. You can go, just go to custom experiments, and then you can disable only that JavaScript. And then you run the test, and then you see that again with the baseline, the comparison of that with the baseline.

[00:35:53] Nathan Wrigley: Okay, so that was going to be my next question. Can, let's say for example, there's, I don't know, 15 JavaScript files and 28 different CSS files. I was going to ask if it was possible to just pinpoint some of them and say, okay, let's disable 14 and just leave this one CSS file here. Let's just see what happens. And it sounds like that's possible.

Does the tool automatically run through a cycle like that, or do I have to, do I have to go in and set up these experiments in such a way that, okay, let's try that one now, let's tick, tick, tick, tick, I want those ones on. I want those ones off.

Do I have to do it manually or is the way of just getting web auditor to slowly but surely over the course of a few days maybe, I don't know, go through and just try a whole different arrangement of things that are toggled on and things that are toggled off.

[00:36:41] Pradeep Sonawane: So basically by default we are doing these experiments whenever user audits any site. We run these experiments just like the baseline audit.

[00:36:51] Nathan Wrigley: Okay.

[00:36:52] Pradeep Sonawane: And then under a minute, all these experiments are performed, and then we basically see the Lighthouse report.

The custom experiments that I talked about is actually after you get the report, you basically have an option, extra option, to basically just run custom experiments. So that you have to run manually afterwards, but, these experiments, removing JavaScript, removing third party resources, removing render blocking resources, removing low code coverage resources, all these are, know included by default.

[00:37:26] Nathan Wrigley: So you run the experiment to begin with. Then you'll get some information within, I think you said under a minute, something like that. So very quickly you'll get some feedback, but that then that feedback leads you down a path of, okay, let's try these things. So you're going to be spending a bit of time in the tool.

It's not like you can click one button, walk away, come back in a couple of hours time and it will have just tried absolutely everything. You have to be engaged with the tool and toggle things off and toggle things off. And I guess in all honesty, that's probably the best way of doing it, 'cause you can keep your eye on what's going on.

Yeah.

[00:37:59] Pradeep Sonawane: So right now, I also felt that thanks to Hotjar tool basically, I would like to take that name because I use that tool to find out the.

[00:38:08] Nathan Wrigley: Is that like a heat map tool? You can see where the mouse has ended up?

[00:38:11] Pradeep Sonawane: It also shows you the user's interactivity at this time.

[00:38:15] Nathan Wrigley: Oh, that's right.

[00:38:16] Pradeep Sonawane: So the recording of everything user does. But anonymized, in anonymized way. So you don't know who is doing what, or you don't know the details, but you know where user is going. Where user is spending time. Where user is bouncing off. So it's from that info I got to know that user is not spending much time on this, on the tool. So it's like user is actually waiting for one minute, and it's one minute is also very long time for the user.

It's what I got from that tool. What I did was, in our next release, we have progressive reporting. It means the baseline performance is actually reported very fast. It's like only 20 seconds. So in the 20 seconds you get the baseline report, you get the performance analysis. You get the report of that.

But then we also continue the reports of the other experiments in the background. We show the loading for that, loading activity. The log off that, basically what is running and what is not. While that runs, user can go through the performance analysis that we have, without AI.

And then when we have those experiments run, when we have the comparison also built, and reported by the tool, we then also have, at the last moment, we basically have AI callbacks. As in, we send the input to AI. Okay, these are my, issues, and then it responds with the solutions. Those solutions we basically categorize between developers and managers. Like for managers, we show it differently. For developers, we show it differently so that the thing that they are interested in, they can look at it.

And this happens progressively so user will not have to wait for one minute, like looking at the screen with a loading icon and all. That's a learning that I took from Hotjar.

[00:40:16] Nathan Wrigley: So it sounded from what you've just said as if there's like a team, capabilities, different roles built into this.

For example, you might want certain data to be visible by, I don't know, the development team who obviously really need to get into the weeds of this particular thing. But maybe, I don't know, the marketing team or something else might have a really different experience. Did I pick that up? Is that actually built into the tool or did, I just guess from what you said?

[00:40:42] Pradeep Sonawane: It's happening at the moment. It's happening at the moment, as in you can see what a manager would need to see, right? You can see the quick wins. You can see the long-term, plan, short-term plan, for the managers. So a manager would know what he has to work on, like what things are to be worked upon, for a short term goal and for the long term goal.

And there is a kind of estimates also for those issues. We can see the number of hours that might take to fix that issue. And so these things are there. Developer would see all the things basically, and managers will also see all the things. But I'm just saying that managers or the owners, or let's say non-tech people, would be more interested in knowing how much time it's going to take. What are the quick wins. What are the low hanging fruits. Those things.

Then they can forward that information, to the developer or to the team. And they know beforehand, like how much time it might take approximately. And then they can expect that thing.

[00:41:50] Nathan Wrigley: Just a quick question, slightly moving gear a little bit. If, for example, I've built a website and I've handed it over to my clients and for some reason I've given the client the capability to install new plugins and off they go. They have a great time installing just 300 additional plugins and the performance of my website has gone through the floor.

Does web auditor allow me to just continually monitor what's going on and then perhaps, I don't know, send me a, an email saying, wait, something has gone terribly wrong here. Overnight we've gone from 300 milliseconds to four and a half seconds. There's something that needs to be looked at. In other words, I'm asking, does it almost act like an uptime monitor service, pinging you when a modification has been made, be it for better or for the worse, maybe.

[00:42:37] Pradeep Sonawane: Not exactly like that, but we have a recurring audit feature in which you can set the URL and you can set the frequency of that recurring audit. So let's say I don't make changes quite often. I just make changes weekly, let's say. So I'll set weekly audit, recurring audit for a particular URL for my important URLs, for that sake, right?

I have maybe let's, say five important pages, and I want all of them tracked and monitored. I will get, if I said that, then I'll get email notification every time that runs every week. And then I can see, I can go to that report, I can see my past audits also, and I can see the trend, whether the performance is going up or going down.

[00:43:26] Nathan Wrigley: Nice.

[00:43:28] Pradeep Sonawane: Yeah, that, that feature is already there.

[00:43:30] Nathan Wrigley: So it's a bit like setting up like a cron in a way. You just say once a week, what have you, and then give me a breakdown of what's happened. Then in the back end, you can chart that over time. So you can see that weekly task, here it is, here's, a month ago, three weeks ago, two weeks ago, one week ago, and here's the trend that's coming out of that data.

And again, this is maybe where the AI drops in and starts to say, we can see that things are getting better and better, well done. Or things are not going as planned.

It also, again, changing tack a little bit, it also feels like there's a really nice piece here where an agency owner could use this tool to add value to their agency by going to the client and showing, look what we did. You didn't ask us to, but over the course of the weekend, we've managed to speed up this page, and this page by, I don't know, of a second or what have you.

Is there anything built into the tool which enables agency owners, I don't know, freelancers, whatever, so that they can report to a client? I dunno whether that would be white label or a particular type of report which might be even more straightforward for a non-technical user to understand. Any of those kind of things?

[00:44:42] Pradeep Sonawane: Yeah, the same thing I was also thinking about. We have already built an extension, the browser extension, and then we also have APIs that people can use in their workflows and then they can try making use of it in their workflow. So they don't have to come to webauditor.io to run the audit every time, but they can maybe just use an API whenever they want to run an audit.

It's there, but it's it's still in development. We are already thinking about that and I would actually want people to use this in their workflows, so that they can basically make use of it and they can also give us more feedback. Maybe we can improve upon that, and it's just an opportunity to improve.

[00:45:32] Nathan Wrigley: Yeah, it's a really great kind of virtuous cycle. If you can provide a tool which the agency owner can then demonstrate to the client, we did good work for you. It makes the client happy. Which in turn presumably makes the agency happy. Which in turn makes you happy because people will keep coming back using web auditor, if they're an agency and they realize the benefit of it. And they'll be told there's a benefit because the clients will tell them that there's a benefit and so on. So it's this nice virtuous cycle.

Okay, so we're probably running out of time. So I've mentioned the URL lots of times, but I'll do it one more time. It's webauditor.io. You can find it over there. But, importantly, I'm going to quickly hit on the pricing tab, 'cause that's a very important part of the whole jigsaw puzzle.

Do you want to just run us through the pricing. Caveat emptor, like I said, we're recording this September, 2025. No doubt the pricing is subject to change. It looks like you've got a, you're binding the pricing to the amount of experiments that you can run and things like that. Just run us through what you've got.

[00:46:36] Pradeep Sonawane: So by default, like people can try the tool for free. The two audits they're getting for free. And if they want to perform more audits. They can basically go to the pricing and purchase one of these plans. They start from $9 per month. If they need more audits to be done every month they can choose higher plan. Teams access also, like they can add team members, their own team members. They can run more audits. The higher plan, they choose, they get more stuff to do. But AI analysis is available in every plan that we have here.

[00:47:13] Nathan Wrigley: Yeah, so start a plan is $9 a month and at the moment, that allows you 10 audits and a couple, two, custom experiments with two recurring audits. And then those numbers multiply as you go up. $19 plan, $29 plan, and then there is a 49 and a $99 plan as well, which is the great granddaddy of your service at the moment.

Do you allow a la carte pricing? So let's say for example, I'm on the starter plan and I, want to have 11 audits, which is just slightly over 10. Do I need to move up to the bronze tier or can I purchase, I just want this one-off just then a single audit this month, which I'm probably not going to have to repeat again. Do you do that kind of thing?

[00:47:54] Pradeep Sonawane: Yes. That's in our roadmap, actually. It's not right now, but we are thinking of adding that also.

[00:48:00] Nathan Wrigley: Yeah.

[00:48:00] Pradeep Sonawane: I'm thinking if it's a rare thing, but maybe that's a possibility. Somebody would like one audit.

[00:48:07] Nathan Wrigley: Yeah, you never know because if you're on the gold plan at $49 and you really have just maxed out on your a hundred audits and you need five more, maybe it's a bit of an edge case but it'd be interesting to see how that goes.

[00:48:18] Pradeep Sonawane: So right now what happens is you can upgrade the plan, and you only pay the difference. You have a previous plan and you like halfway through it, you still have more audits, but you want, want to make sure that you have sufficient audits to be done this

[00:48:36] Nathan Wrigley: Yeah.

[00:48:36] Pradeep Sonawane: You can upgrade the plan. You pay the difference.

[00:48:39] Nathan Wrigley: Oh, that makes sense. Yeah. Thank you. Okay, I think we'll call it a day there. Hopefully we've given you some indication, dear listener, of the capabilities of webauditor.io. Go and check it out. Is there anywhere apart from that URL, is there anywhere where you hang out? Let's say for example, somebody's listened to this podcast and they've got questions for you specifically about this product. Is there any place that we can find you? And if you say something, I'll make sure it goes into the show notes.

[00:49:04] Pradeep Sonawane: Yes, I think, you can basically contact me from the website. Also there is a contact form. It'll notify me. But if you want to connect with me, maybe you can contact me at my email address, pradeep at rtcamp dot com.

[00:49:19] Nathan Wrigley: Perfect.

[00:49:19] Pradeep Sonawane: Then I have a LinkedIn account also, so you can hit me up on LinkedIn.

[00:49:24] Nathan Wrigley: Lovely, lovely, lovely. Thank you so much. Pradeep, it's been absolutely a pleasure chatting to you today. Hopefully we've managed to persuade the listener that, it's worth a look. webauditor.io. Go check it out, and thanks for chatting to me today. I really appreciate it.

[00:49:38] Pradeep Sonawane: Thanks. it's been a pleasure actually. It's been wonderful talking to you.

[00:49:42] Nathan Wrigley: Yeah. Thank you so much.

Okay, that's all I've got for you today. I hope that you enjoyed that. If you did, head to wpbuilds.com. Search for episode number 439. Leave us a comment there. Please, please, please. I love it when I get comments on the website. That was one of the things that made WordPress so fun back in the day. I know that we've kind of resigned ourselves to posting things on X and Facebook and all over the place, but it was so nice when somebody drops a comment. wpbuilds.com, search for episode number 439.

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 a lot 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.

We're also helped out this week 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 are outta this world. Experience it today at bluehost.com/cloud.

And we're also joined by Omnisend. 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? Well try Omnisend today at omnisend.com.

And dear listener, we would like to thank GoDaddy Pro, Bluehost and Omnisend for their support of the WP Builds podcast. If you would like, as they did, to get your message in front of our audience, head to wpbuilds.com/advertised to find out more.

Okay. That's all I've got for you today. Well, very nearly. Just a quick reminder, we have our This Week in WordPress show. It's at wpbuilds.com/live, 2:00 PM every Monday. Hopefully, you can come and join a whole phalanx of people who drop in the comments and make that show entertaining. I'll package that up as a podcast episode, and we'll see you next week for a Thursday podcast episode.

You stay safe. Have a good week. Bye-bye for now.