Build an SEO & AI Visibility dashboard with (mostly) Free Ahrefs APIs

Okay, let's go ahead and get started. Hello. Hello everyone to who is joining the webinar. Shout out to where you guys are logging in from and uh yeah, shout out to where you guys are logging in from for today's webinar. in the chat. Hello. Yes. Yeah. To everyone who's joining in, welcome. Just going to go ahead and quickly share my screen here. Awesome. How are you guys doing today? Hope you guys are looking forward to today's webinar uh live session. We are going to have a lot more uh vibe coding AI focused stuff coming on. Uh is everyone hearing an echo? Is it just me? I have echo. No echo. Okay. You may have the uh live session open twice on two um on two browsers. Just check to make sure that you only have it open once. Yeah. All right. Cool. So um how we're going to do today's session is that I'm first going to go through uh the content and show you guys what I've built up and then we will go into the questions. So you can ask questions while we're going along. Just make sure to know that we will be answering them at the end of today's uh live session. uh that is during the time and we'll be answering them at the 30 minute mark. So like right now it's around five. So uh we'll first go through what I've prepared, show you guys what the dashboard looks like, how I've used the HF's APIs to build them and how they can the data becomes useful for you guys to improve your AI and search uh performance. Then we'll go through the questions. I will also take that time to go through the best practices or the things I've learned while building this uh things that you can take on to build your own tools and dashboards and save the pain that I went through and not repeat mistakes. Does that sound good for you guys? And if that is good, then uh let's go ahead and get started here. Um today's webinar idea actually came about when I was researching uh our brand visibility using brand radar as well as looking into like what we were ranking for organically the questions people are asking and a lot of times AI uh generated answers or Reddit forums would say yeah HFS is very expensive and uh you should consider the these cheaper tools um and plus on on another level since we're all like vibe coding and trying to use AI to create like custom dashboards or like automating workflows for our needs. Um the side problem is that we have to worry about like API costs and how many tokens we're spending and um that becomes that starts to add up the more we start using AI the more we start using API and then then it becomes a bit of a worry. So really today's webinar is a bit of a challenge where I say can I build a dashboard that you guys can use with your atrif subscription that uses the endpoints on our platform that cost zero units to pull data and actually there is a surprising amount of information that you can get and for the longest time accessing our data via API directly was limited only to our enterprise plan. Well, that is no longer true. Light plans have API access. Light plus plans have API access. Now, um based on a plan that you can get, uh you get more rows per um request uh as well as total units that you can spend per month. But guess what? because of today's webinar and like um what API endpoints you're using, you don't have to worry about this units per month that you need to spend since it's going to be zero. Okay, so that's the plus. You just have to worry about the number of rows per request that you will uh that you'll get from the API. Uh yeah, and so the higher plan you get that the more uh rows per request uh you'll get back. and 10 seems like a low amount but uh I will show you some tips about how you can go about that later on. Uh next so let's talk about these free API endpoints. So uh in this dashboard I use the web analytics uh endpoint GSC insights rank tracker and brand radar custom prompts. So the other point about the today's HRS plans is that for if you have a light or higher plan you get free custom prompts that you can track for AI answers in chat GPT perplexity co-pilot and Gemini. So before you set up this uh dashboard that I showcase to you today, make sure that you set up uh your the website that you want to track analytics for uh as a project and add the web analytics snippet uh to collect clicks and and events and stuff like that. You can set up uh keywords to start tracking your rank tracker and set up your uh report and start tracking custom prompts and brand radar. All of these things you can do in your current plan if it's light or higher. So you can set all this up. You don't need to pay additional money to to get this uh information. So and if you look into our documentation, you will see that endpoints will have this notification to show that it does not consume any API units. So that is useful to look out to when you check through our documentation if you like to tweak uh or or build additional widgets based on what we showcase today. Yeah. And um it is very important to like check out the documentation even if we have AI because it really helps to make sure that you have exactly the data that you need in your dashboard when you um build build this dashboard out. So I will talk about what so I mentioned that this dashboard has two pages. So I'll talk about what data it will show like I have structured it out for each page and what it's showing uh as well as a bit of information about like why um I have split it out this way. So the first page of the dashboard goes into performance. So I want to see uh what is the share of voice that I have both in search results. So like things like uh how much how often is uh results of my pages showing up in Google and of course we estimate based on like the number of page the results that are from my website versus my competitors and from that we calculate a share of voice. Uh we look at impressions. So even if people don't click, how often do I show up uh for keywords and like and how often are those keywords searched? Those are calculated into the number impressions. So if my website is showing up more often for uh keywords that have are searched more often, then I have more impressions. Then I can look at how that compares to my competitors as well. And then clicks. Clicks are still like useful. It helps me get engaged of like what uh pages are getting um clicked uh like are getting visited by people. Um in in future uh live sessions I can talk about how we can calculate the clicks that are coming or the visits from bots but today we'll focus mainly on like human visitors that are visiting your website. It gives you an idea about from not only like how how much clicks you're getting per month but also like the sources uh and top pages. So I still want to see what are the performing like which are the pages that drive the most uh clicks u whatever the source it may be. So that's page one and these are like um like at least for clicks and and like uh top pages those are familiar to you. uh they're not so super fancy. Uh what I'm really excited to show you guys is dashboard page two which talk about opportunities. So uh it shows questions that people ask about your brand either through uh keywords that you rank for organically but also uh questions that get asked in people also ask uh widgets uh for the keywords that you rank for. Um we will also be this dashboard also show uh discussion and forum threads like Reddit and Quora for the keywords that you rank for. And these are um places where people are getting um topics um they're getting their questions answered. These are places where people are discussing either about your brand or uh topics related to your brand that you may consider either participating in or uh creating content around. uh most cited third-party pages and domains. Um third-party brand mansions is extremely important in influencing um where and how often your brand gets cited in AI answers. We have a lot of data studies that we've published about how important it is to get cited on these places. So there will be a handy widget that basically shows the top widgets I mean the top uh party third party pages and domains that uh you rank for or rather that uh that site your brand. Okay. And also there'll be a widget that helps you to show and that showcases videos uh that rank for the keywords that you're targeting as well as get cited in uh AI overviews and other uh AI platforms. And why? Because not all mentions are equal. A lot of these uh major um websites like YouTube, Wikipedia, Reddit, major publications, they tend to be used more often in AI systems. they're trusted more often. I have added the links here at the bottom and you will get a copy of today's slide. So, you'll get to click and see our studies on this uh that that show how how much more often these like um third-party websites uh are chosen are are preferred by AI systems to generate answers from. So it's it's good to know what these pages are and and what like these pages are talking about so that you can create your content more strategically. Can you show page one for a second? I miss it's just this share of voice impressions uh clicks and top pages. Okay, that's your one second. Going back now let's move along. I will show you uh more information or rather not show you but I after today since you guys have joined live I will um send you guys like a packet that goes into more details to how I prompted cloud code but in general this is my this was my approach. I gave an overview of what I wanted my dashboard uh to show. I tried to be uh explicit in showing which data points from which sources that I want to source from to create that data. That's really important. Um I told the AI to use and reference uh HF's MCP connector to make sure that they have they have like a proper uh schema to reference whenever making these API calls. And I also supplemented each of these widgets with this curl request. So what is this? So if I do I have this open uh but yeah on your platform uh in like uh site explorer or web analytics or JSC insights you'll see this like API button right this API button that's on the side and like it will generate an API um URL based on the page you have open so you can like filter for uh a spec specific location or keywords that are three or longer. or have certain filters applied and this will update the URL. So you what you see in that report is what you get in that API URL and then you just copy that and reference and you know put save it in your uh instructions document that you give to AI and say like okay I want this this is what I want these are the parameters I want to apply and that becomes really important for them to site so that they don't like mess this up uh too much because if you the more explicit you are the less likely they're going to hallucinate things and or like sometimes I find when I was debugging why like widgets were not working or the data was wrong was that it would just decide to put in random stuff thinking that the API can can apply it you avoid this so much more by giving the specific like uh API URL here and our platform helps you build that this is also your chance to basically understand what is actually possible uh to pull out of uh like what is HF's data that you can pull out because you see it then you can take the URL and then you can give it to your AI to vive code whatever you need to vive code. Okay, so let's get into it and I'll show you guys what I've built out. Um so what I'll do is for each widget I'll explain what data it shows, where it gets sourced from and how that would help me or you as the user to um like make decisions. So let's get into it. So this is what I built over here. So with each of these widgets uh I added I asked AI to basically give this API button here to show what endpoints they used and what parameters they used and what they calculated after pulling the data. So that gives me an idea whenever I look at something like this like the the charts that show up I get a good idea of u what actually has happened to show me um that to to create the the results that it shows me in each widget. And this is really helpful for debugging or checking if something looks wrong. It also gives me an understanding like sometimes I don't remember what uh logic I've used to to create this. I may not remember what I'm looking at. So this is really helpful for me. Here it's taking uh brand radar um share of voice history. So it's able to because I've added in my brand radar report my brand and a bunch of competitors. I can see and and and I'm tracking a bunch of like custom prompts. I get to see over time how that share of voice um it compares against each other, right? And so that gives me an idea of like how often I'm I'm showing up uh in AI uh responses compared to my competitors for the prompts that I'm tracking. I hope that this increases over time. And then like if I see that there is a sudden increase from my competitor, I can then go and investigate based on like the date here. um like and then go back to the HS report and pull out like the the prompts to see which ones they have suddenly increased um mentions uh number of mentions for. So that's how I use the data here. Okay. I then do the same thing for organic search. So brand radar doesn't have information about uh how your website performs organic search. That is uh rank tracker. It is also um yeah it is mainly in this case rank tracker. I want to compare it against my uh competitors. The API endpoint is this one here where you can compare competitors. And what I really like about rank tracker is that uh it will because you are collecting um SERs for the keywords that you're ranking for. I can pull out the exact SERs uh search results for all the keywords I'm ranking for and do like additional analysis that you'll see later. Uh so again the same thing I'm looking at the overall changes uh and and like the leaderboard of like how much uh I am comp like my share of voices compared to my competitors here and if there's any like uh changes in uh the order or like the sudden spikes I can go to that specific date or check out that competitor's uh like top keywords and see like where they're getting additional uh websites that are ranking or which keywords they have suddenly ranked for uh more often than mine. Uh yes, a recording will be shared to uh everyone who attended as well as people who did not attend but for people not attend it will expire after a while. So you guys are lucky for joining today. Okay. Uh impressions. Impressions is as I mentioned earlier the like how often your brand shows up uh for the keywords that I think in this case this is AI overviews. Yeah. So no no this is not AI overviews. Of course not. Um this is for keywords. This is impressions for uh for like how like where your brand shows up in AI answers based on uh how often actually no this is AI overviews isn't it? Yes. Okay. Sorry. I I'll check in this, but like I I want to look at how how often um how much more often I show up uh compared to my competitor uh for a different in this case it doesn't compare to competitor. It's just my my brand and I want to see like how much more um of the popular queries that show up more often uh for each uh platform. So this is what data is being showed right now. So, and this is important for me because I want to see um how I am trending across different platforms. And the reason why it like only starts after a certain date is because I only added the custom prompts after a certain date. I I did not add it like one month ago. And that's what happened. Um if you change the date, does it resend out API requests? It uh there are certain API requests that give you data per week. then it will it will take from the previous API pool that already has that like weekly data and there are uh certain API requests that like you need to call every time to to change the date every time that it will um cause an additional API request but of course we don't care because uh no units are spent so we can call as many as we want and like um for for the data that we have but yeah uh that's that's how it works I want to know which are platforms that I show up more often. Maybe like there are certain platforms that I care about more often. I want to see how they grow over time. So this is what uh this dashboard helps me see. And then I can see this uh for organic results. So here I can see uh for the keywords that I rank for, not necessarily uh that people click on, but if I show up for example in a top 10 results or top 20 results, I can I then see uh for which keywords and how often those keywords get searched per month. So if I show up for keywords that uh have more searches per month, my impressions go up, right? And I can see that uh from GSSE data. So because I've connected my GSC uh to HFS, I can pull that information out and see how that uh ranks over time. And and similarly from GSC I can see uh clicks and here just clicks alone is not that like uh you know actionable information. What I also want to see is um which subfolder uh drives uh has the most uh clicks going on. In this case, it's very obvious uh that English has the highest number of clicks as you can imagine. Uh but you can set up the subfolder to be whatever your website is set up like for blogs or certain articles or article categories or um different uh sides of your website and use that to sort uh see how much clicks and impressions that you're getting um for each of these sections and that becomes important in understanding which are the kind of strategies assuming these subfolders uh refer to specific uh strategies or or areas of your website of interest and you can see that if that changes over time and uh how like which one of these categories are or subfolders are driving uh more attention organically. Uh and there's a number of questions here. So like I I did mention I'll answer them at the end because I want to make sure we go through everything first before we answer them afterwards. Okay. Clicks over time. So um we have this uh special feature in web analytics where we automatically categorize clicks from AI uh platform. So by clicks from AI platforms I mean when people uh search you know whatever they do on AI and then AI generates an answer and they site sources and when they site the source some of your pages show up as a source and and these people click on that uh source within the AI answer and there's you sometimes but not always it depends on what AI does a referral code uh in a URL that tells uh tells me or tells web analytics that this traffic, this click came from a cited AI answer on these platforms. So I can set up my widget to see that over time and that matters because like um you can see your pages getting cited at AI answers but this is actually proof that people are clicking on it. There's interest to click uh on the cited sources in AI answers. you can go into additional detail to see like what are the top pages that get clicked on by um like uh uh that come from these AI uh answer sources. Uh but this is I just want to show you you guys that you can create a widget like this this way and that becomes useful for me to see like which are the platforms that are driving uh more clicks than others. In this case like yeah chat GBT typically gets a lot more uh than the others. Okay. So now we look at top pages. So uh what I can do here so even though we say that in brand radar you get to see what generate what answers are generated by AI as well as the cited sources. What you can actually see today on your plan even if you haven't bought the the add-on is what pages are cited in AI overviews. So I configured this widget to to fetch from rank tracker um the whenever there's a keyword where AI overview shows up and my website gets cited as an answer I wanted to pull up those top keywords uh those top pages here and this is what this widget's about. So I can see without you know with like on my current plan with rank tracker what pages get cited in rank tracker for the keywords that I'm tracking for. So this is what um I see here as the results and yeah so that lets me know what uh AI is or AI overviews is picking from my pages to site as an answer more often than not. Uh here I'm looking just this is just straight up like looking at top organic pages by clicks and by impressions and that just tells me uh which are the the pages like this is your standard if which are the pages that are driving more clicks organically or showing up more often in uh AI in in uh Google search results and that's like um I can then configure this to to check for like the ones performing the least or they have any drops. um between the dates and investigate to see if there was any drop in uh positions or maybe there has been a new competitor that I should be worried about and that is something that I can pull uh using using GSSE. Okay. So this is from GSSE uh insight. So and GSC is directly like this is directly from Google. So then I don't have to worry about like discrepancies in data per se. This is what uh like compared to datas that is being estimated uh at the HFS platform. Typically it this is going to be more uh accurate. So pick that instead. Um so now let's look at page two the opportunities. So if we go back I want to show you like I as as we referenced uh about what page two is all about. I'm looking for questions people ask uh where uh like discussion forums and threads that get cited for keywords that I I'm I'm interested in that's important to me. I'm looking for third party pages and domains that get like um that get cited in the AI uh custom prompts that I'm tracking and I want to look at the videos that get ranked and cited for the keywords and prompts that I'm interested in. Why? because these are like the um like formats or content ideas or the uh creatives that are going to be showing up as answers or influencing the knowledge that my um target audience is receiving as they get answers from both search and AI. So here we have uh the endpoints use is cited domains. So, because I've entered in a bunch of custom prompts and in those uh and I'm tracking their answers on these chat, GPT, Gemini, Perplexity, and Copilot, um they site uh sources. So, I want to see what are the top sources uh based on what they have uh that I get cited for the keywords that I'm worried about. And this is like a list of um places here. So of course like it's good to see that uh my own website shows up but then I also see that there are this u this major publication for SEO search engine journal is there I get to see that LinkedIn post show up a lot YouTube shows up a lot so this is like useful information to know uh what what are the platforms that get cited more often by AI and I should definitely check to see how my uh brands how my brands are being referenced on these platforms. or if they're not, then I should really consider getting onto them because this also includes uh custom prompts and tracking against my competitors. So, um if my competitors show up uh in these um domains, but I don't, then I really have to consider um checking them out. Okay. I've also set up uh thirdparty URLs and air overviews. So again as I mentioned uh earlier that you can source the URLs that get cited in AI overviews um from search results within HFS. So here it is taking it's done a few steps here. So what it does is that it first goes into rec track and said please filter a list of keywords where AI overview shows up and then for each of these keywords where AI overview shows up please pull out use SER overview as an endpoint to pull out uh all the third party uh URLs that get that get cited and then rank them for like the number of times I want to see which URLs get cited mo most often. So this is what this widget shows me. So I can pull that information out straight from rank tracker itself. It does calculations in a few steps, but these are basically ideas for content and questions I need to answer in order to show up more often for the uh people that uh for for the keywords that I'm focused on ranking for. Right. AI overviews rank tracker. Okay. Um, Alex asks, "Is there a step-by-step process to creating this dashboard?" Uh, uh, like I'm trying to show you guys like why like how I've structured this dashboard and and how I've prompted it. Uh, you'll get the actual prompts itself that you can use to build it as well as the dashboard itself. Um, tomorrow in the emails sent out to everyone that has attended today. So, look out for that. Okay. Question keywords. This is um Google Search Console uh your keywords but filtered for any keywords that have like what, how, where, why, who, uh when sort of questions. These are questions that um these are questions that people typically would type not only in search, but they could also be typing that into AI. And so I need to know what these questions are uh and then see whether or not I'm answering them anywhere on my on my website or in other people's websites. It's good to to do an audit like this. But it's important to know what people are asking. So um that's why I have pulled out this information here using GSC. Okay. So you can see that it does a filter like this. I've also filtered out for uh longtail keywords because they're also kind of like question keywords or like topics and people might be typing in. They're also more likely to be uh related to things type people type into uh AI platforms. So in this case you get to see things like people are searching for AI shared of voice uh problems that they face on a platform. Do I have answers for that in my documentation? um how do I am I am I putting out enough information about how to solve particular problems on our platform right again this is GSSE you can see here right okay uh people also ask top questions ah now this is an interesting widget that also does the same thing as earlier where it first goes into rank tracker and then it pulls out all the keywords that al that have this people also ask SER feature And then for each one of those keywords that have P people also ask show up, it pulls out the search results and filters out only the questions that people ask here. So this is like when this um when this uh uh people also ask show up in search results, these are obviously created by Google and they are like searchbacked ways of finding questions that people typically ask about the keywords that you're trying to target. So the question is are you answering any of these questions or well enough on the content you currently have? Uh, and this is a way to pull all this information out immediately um through uh the tools you're already using to track uh your your your website for. And fun fact, whenever we populate our brand radar main database and we uh we actually do exactly this, we pull out questions from uh people also ask uh with uh SER features that show up for uh the keywords that uh that people search into Google and we use that to ask these questions into uh like any AI platform like chat GPT and Plexity to see what is what gets answered. So this actually this is a direct source of like a lot of the questions and prompts that we have in the main brand radar database. So you're just doing a portion of that uh using rank tracker which is cool. And then uh similarly I believe this is also rank tracker. Yes, it's doing the same thing but now for discussion and forums feature. So now it's pulling Quora, Reddit, uh different posts, uh local search, forum posts, and these are all the URLs that show up for that particular search feature. These are again places where people are discussing either about your brand or things related that could be related to your industry as well as questions you could possibly answer uh in your content. They're all things that these are all ways to to cover. And of course, these get cited more often in AI answers as well. So you really have to check to see how you are stacking up um as your brand being mentioned and how it's being mentioned on any of these uh pages. Okay. So again this is pulled using a rank tracker overview the same way filtered for discussion and forum search feature and then you pull out the search results for it and you filter for only those search features and get the URLs. And here we have the keywords that it that it that it fetched from. So that's good. uh here now this I'm taking cited now this is you know it's being cited in AI answers already because this data is being pulled from brand radar from your custom prompts so if I filtered it through for just the Reddit pages uh this gives me an idea of what is being discussed and being cited as um AI uh answers again it's like same thing but just a similar source uh where I'm looking for topics as as well as places that um my website that that that I should potentially get involved in if if I want to for like in these uh discussion and forum platforms. Okay. Uh popular video topics as we mentioned earlier uh that videos get cited a lot like YouTube gets cited a lot as answers and they also show up a lot for for answers now uh just in search normally. So, it's important to pull them out to see what they're talking about. And like currently with the data that you have, I believe this comes from rank tracker. So, you see, you can really pull out a lot of interesting uh information just from the tools you already have. These are um topics like tutorials about different tools uh or to go about a specific workflow about specific uh like metrics on our platform and stuff like that, questions about our data. So, like these are these are uh websites that I can take in or uh that I can check out and see like what they're talking about. Who are the creators? If they show up more often than not, maybe I can consider working with them. Uh as well as what kind of formats that they take. Are they long videos? Are they podcasts? Are they, you know, um shorts? Here the shorts is showing up. Okay. So, yeah. And then the same thing is happening for AI uh for brand radar custom prompts. So I've asked I've extracted out the uh YouTube links that are showing up in the AI an AI custom prompts that I'm tracking and I'm doing the same thing. I'm looking into the creators, the formats, the topics that are being mentioned uh in these results here. Okay. But this one is specific to brand radar cited pages. Okay. All right. So this is like a very quick overview of like the kind of data that is available to you already with your HS platform. You just have to configure it out in this way and uh in in these ways to show you that information that is useful. Um because this is a 30 minute webinar. I will now go straight into and we're actually we're already slightly over time. I want to go back into our slides and just talk about some general tips on vibe coding uh these dashboards and using the HF's API as well as MCP. The first is you should start with one widget. Um, if you're especially if you're not too familiar with how the like the API works or the documentation or what data is available, start with one widget and start building one after another because once you get one right, then you have an a rough idea as well as a template for AI to use to create the next one. Okay. Um I although I did create a dashboard um using like a like an overall approach. I used this basically as a as a context outline to what and telling a that I want a lot of like these things to be on my dashboard but I still gave the explicit like API calls that I was looking for right and I used that from from the documentation. So, but still if you're not sure, you should start uh one at a time and you should you can ask API an AI to say please check HF's MCP for information because the MCP has um if you have it connected with your AI has information about the schema of the available API endpoints um and what sort of data comes back so that it knows which endpoint to use in order to if you're looking to build a chart or just a table of results or what like data is is possible. possible what sort of fields are required in order to pull that information because that will help you save so much time in like uh data hallucination of of of AI thinking that uh you know you should add this field or this parameter uh I've used the HF's API documentation page so on this on all our documentation pages we have this like so I gave an example here this widget on the side here and what you can do is you and um use this to put in like information to test what results you should be getting out like you just put in the API key and like fill out the required fields and not only will you get the URL that you need in order to pull the data that you're looking for, you also can check the responses to make sure that it matches uh what you expect it to come out. And this is important because that sometimes um you need to check whether or not the the numbers or the results match up or sometimes you may see a bug and you want to compare it. You have a very easy way to do it here. Okay. Uh and then tell AI to handle errors and exceptions. Sometimes uh the widget just shows no data and you have no idea. Sometimes a widget will show like zero or like you know a thousand or 10,000 and you go okay that doesn't seem to match up with what I know this widget is supposed to show and uh if unless you don't tell AI to like um show errors when it faces errors along different steps uh it will not tell you. So a lot of times it depends on on like how smart you've configured your AI but out of the box if you create your own dashboard and like if you don't tell it ahead of time to handle errors and exceptions it's not very easy. So I've used that. It was super helpful to figure out whenever I I faced issues and then I used those error messages and copied it back into AI and asked to say hey you know I got this error information can you please help to investigate and again that's also where it is helpful to like cross reference with like try triing out a response here and then seeing what response AI managed to get when it vibe coded something. Okay. Um so this is I want to do a quick promo for um our two conferences that showing up this year where we will talk about like you know what's happening in the world of digital marketing as but also like plenty of speakers that uh cover topics around AI AI visibility as well as search and automation for uh marketers. So, please, if you like to meet a lot of us and that are attending today in person and and more, uh please go check out our conferences and see if and consider uh buying a ticket and joining us. Okay, so I'm just going to do a quick shout out here and that's a wrap. So, like there's going to be a bunch of questions. So, I'm going to answer the questions. I think I'll first answer the questions in chat and then I'll answer. I think people have been um putting them in in the Q&A. Okay. So, first I'll answer the questions in chat here. Um all the way up all the way all the way. Now is also your time if you have some questions to put them in the Q&A uh section there and I will help to look at them. Uh is there is the share of voice Amelia? Sorry if I butcher your name. Ameilia I think asked is share of voice taking into consideration only organic positions or is it taking into consideration Google features and Google ads? So um share voice depends on which source you're taking from uh because it calculates share voice slightly differently. This one here I um this is I believe we have an article that goes into detail about what this metric means and how it's calculated. Also um if you hover on share of voice uh I think because this question is talking about organic share of voice it um it is talking about rank tracker. So in rank tracker it looks into the total clicks that your target receives for all tracked keywords and versus how many clicks the the search results is receiving in general for that for that keyword. Certain like SER features we get information about clicks certain SER features we don't. For example, there are certain search features like knowledge panel that has no links uh that come back to us and so we can't really estimate like how often it gets clicked. it's not like really doable and there are certain sort features where we do get information about clicks and we're able to estimate clicks and if we do we do so so but it still follows the logic here calculated just know that share of voice is calculated differently I think uh on uh on brand radar because that that is looking into more like mentions and proportion of mentions of your brand versus like other brands in AI answers things like that. So just keep that in mind. But it's helpful because usually uh on the platform if you hover over share of voice we have information like just hover over the header you can see uh how it is being calculated there. Uh on I think Hun asks can you compare the number in this dashboard with the actual number in HS please? I tried MCP before and the number is not the same. Uh, I'm not sure if I can quickly do this live actually, but um there is sometimes it it does not match up and and that can happen because AI does weird things. What's a good What's a good one to see? GSC maybe. Uh so yeah, I I do have to go into it sometimes or sometimes I see that keywords uh that should show up as as uh results that don't show up and those are also things that that need de debugging that can happen. So let's compare it to just make sure that your filters match exactly. Yeah. So, and you have to make sure that you're comparing last 30 days if you're using that on a platform and things like that. So, I would do this and compare. So, we're looking at around 40k 50k 40k 50k. So, I I will go from day to day and track this. But yeah, I can I can see that like it should like match up roughly just like it's a bit hard to but yeah, this is how you should do it as well. 49 49. Yeah. So, so like yeah, you can do this to to make sure to check if you use um using MCP uh should help you reference the APIs used, but like that's also why especially for when I coded this dashboard, this mainly uses the API directly. it doesn't like pull like data from MCP as an endpoint like and and if I use the API if I ask if I build this using the HS API directly I can then actually see uh what data gets pulled and I can test that um that I find it helps to um reduce a lot of these sort of like weird bugs where you don't know what's uh what what the issue is. Okay, I hope that helps. Um will you share the prompts and APIs that I can use? Yes. uh you will get a copy of this uh like a link to the GitHub repository uh tomorrow when the emails go out. Uh it also gives me some time to make sure I package it properly so that you guys can use it. Uh I have to clear a bunch of stuff like uh it it uses a bunch of global settings as you can see here. Uh and uh I don't want it has to be clear for you to use stuff like that just checks. But yeah, you'll get especially if you guys attended live. So, yay for you guys. Uh, is the air overview API free to use? Muhammad asks. So, let's go to the documentation page to see what which um uh like again to to explain what I've done here. So, rank tracker, specifically rank tracker. So, you set up your rank tracker to track key like your target keywords and its locations and and things like that. Then you can pull overview. So you can what this does is that it returns a bunch of keywords based on your filters. So I'm filtering for keywords in my rank tracker where best position kind is AI overview, right? And uh so let's go into it here and show you right show you guys live how it works. Um so I want to look at help articles here and I see my keywords. All right. So what I'm doing but in API form is adding a filter for cert feature where I show up for AI overview. Okay, that's the first step. And then the second step is that I pull the search results SER overview. So in this documentation, it's this one here, SER overview. Again, it's free and it does not cost any API units. Uh, and AI overview is here and I'm a source. Okay, so I'm pulling this here. I'm pulling this and then putting it into the widget. where is it? No, it was over here, right? Yeah, there you go. See the two pages, two results. And then if I check my uh filter here, it is also two results, right? Data matches up. Okay, so this way of pulling data is free. Okay, Adrian asks, if you change the date, does it recent? Okay, I've answered this question already. Uh, how do I know which prompts to track in AI in Brand Radar? Aha. So, how do I get ideas to uh track custom prompts? Um, if you look into your plan now and you set up some tracking, uh, I'll show you based on the project that I used here. So, first I set up a report where I look there. I set up my website and I set it against competitors and then I tell I set the report so to check how often I'm checking the prompts that I'm tracking and then it tells you the estimated check. So your plan has an comes with a number of free checks per month which means that if you're okay checking less often you can get to track more more prompts. But if there are certain prompts you want to track like more often like maybe it's like live situation or maybe you have uh like a e-commerce website that's very sensitive to specific answers something like that then increase the frequency and it will use up more checks. Okay. Okay. So once you set this up so if I go into this report here there is a setting here for tracked prompts. set up all these like prompts. Did I come up with all of this myself? No, I did not. What I did is I clicked the generate button here. Uh because we quickly noticed that yeah, this is a common problem that like even we are going to struggle if we were to try and find to create all these prompts ourselves. So, but you can use AI to help you and we have a handy tool to help you create these um prompts to uh track uh AI answers for. You can also just go into like your AI tool of choice and say, "Hey, I am you know setting up tracking of answers. I want to look for frequently asked questions of of people. Can you search online or or stuff like that?" Or you can use the dashboard that has already been uh created here for example and look at things like people also ask Right? Put these questions in. You can put these questions in as prompts to track. That's also something you can do. Right? You can put in questions inspired from all these forums and discussion forum pages. These are also like questions you can track uh answers for. All right. These are your ways of finding uh prompts to track in AI uh Uh Muhammad also asked, can we keep this project on local host or should we host it uh on a server? That's up to you guys. um like there are different ways to do it. I'm not going to cover it in today's uh webinar. It it's uh there are a few ways that you can go about this. This is like this particular project or report today is sitting entirely locally on my on my PC. You can host the whole thing on on a server of your of your choice and have it accessible to teammates and stuff like that. And uh yeah. Okay. Um there are some questions people asked are the insights created on the HRS platform or will I need cloud to create them? Uh what did I use to create this particular dashboard and how I pulled um is claude code? How I pulled this data out here is uh using a few steps but entirely HS APIs uh that cost zero units. So, um, they're just pulling from what data already comes from the the the tools and reports you have access to on the platform. I've just reorganized it into widgets that are useful. So, um, they're but they're pulling mainly like websites, website titles, SER uh, SER uh, like fe SER features and and videos and stuff like that. It's not there's not any additional analysis on top. That's and but if you want to do that you can you can vibe code that separately but that doesn't like come from the HS platform specifically. Okay. Uh all right now I'm going to go through the Q&A here. Um how much API credits did you spend for your example of dashboard building? Uh good question. Usually when I build something like this inclusive of debugging, I spend uh maybe a few hundred,000 tokens. Depends. So um you I'm I'm not necessarily looking for efficiency here. I mainly looked into how to get something up and running. you can probably do a better job than me. And now that I'm sending this project to you, you're going to get this uh dashboard up with far less tokens since you already have something working. Uh so consider that a freebie for you guys. But yeah, um I'm not sure how to I'm not the right person to talk about how to get more efficient in a API spending uh for for AI, but I mean in terms of uh using the ATF API is zero units. All this data is zero units. So it's free if you have a light or hire plan to get all this data into a dashboard. Okay. Uh yes, I've already answered this question that Otman asked which is will you share this dashboard with you guys and yes you will because and especially for you guys who have come to join us today in today's uh live workshop and um Jenny asked what kind of format did the output of your example have HTML this is a this is just a web page so it's a web page okay so can I show you guys this maybe uh if you use a AI platform like cloud code for example but there are also like other platforms that have followed uh best practices cloud code has like a plugin or skills that you can access to to help you plan your projects better uh the one I used to help me for this one is superpowers and I asked the first plan um how this uh whole thing will be built and then I checked the plan so uh I have a slide here that shows you uh that shows yeah this one I think here. So, um I asked it to okay, first map out your logic and like any like improvements or or like what each of these widgets and how it's going to build and and what the logic looks like and then I I looked at their plan and I corrected any mistakes and then once that plan looks solid enough that I say okay I can't spot any like obvious errors then I used and I asked it to implement. So first you you can ask the to first like sort out a plan of how they want to implement uh what you're looking for and then uh build it. This gives it a few steps for you to um save time and tokens when building something out. Okay. Um how to make the integration to my project management system so customers are able to see this uh this particular report. I think someone asked a question similar earlier about how whether or not I can host this dashboard and share with more people. Uh you can it's just uh like this whole thing is like basically a project that's sitting in a folder on my laptop. It's entirely local. You can host those files uh on a server uh yourself and there are many ways to to go about it. You can even ask AI and how you can do it. Uh but I'm not going to cover it today in today's webinar about how you can go about these things. However, I will tease that we have this that's coming out. So if you have not heard about uh this new uh platform that we're uh building out for you and other marketers and other people who are interested, we've built our own internal platform that you can use to like create these uh reports and uh projects directly using HF's data without worrying uh worrying about API. Think of it as like a like a application widget building platform entirely focused for marketers and stuff for that and that like would once you've built something on our platform here with agent a you can immediately share that with people in your workspace and and stuff like that. So that's also an option to you. But uh we've only recently pushed this uh particular platform out to uh like people who signed up on the wait list. So I'm not covering this particular tool in detail just yet. Okay. But yeah, um there are plenty of options today really uh about how you can go about these things if you'd like to, you know, bring this um beyond just uh data for your own needs, which is the beauty of today with AI and and uh and hrefs now that you have access to our data directly. Okay. Um we are pretty much at the end here uh of today's uh live session. Are there any more questions? Now is your uh chance to ask. But yeah, so I I have some examples here of like things that AI can do if you that that it can build if you don't know what uh it like how the API is being used or what exactly it returns. It actually is just returning a single number, but it build a line graph. Why did it do that? I don't know. So it's very important the more you know you you understand our API, the better you avoid these mistakes. Um yes, you will get a copy of our slides uh from today for sure. Yeah. And I added these uh helpful examples here to show you like the kind of things that would be helpful to avoid like whenever you get a block pointer uh you find that things are not matching up. This could be useful for you to understand what actually is going wrong. Uh I also have this page here where I I'm showcasing how you can put in the information like the kind of API call you want to make and like what response you'll get. So this is useful because like in API calls you can configure what data you get returned and like put in filters. So, especially if you're on like the lower plans and you don't get many like um responses back uh like many lines of responses back because your plan's limited, you can filter you can add in parameters to filter down your data so that you can like get less results but more like specific to your needs as a response and then have it within your uh rows per response uh uh request limit. Okay. So definitely use the tools available to you across our platform to help you get what you need. And uh you'll find that and that's why I say that you should start with like one widget at a time if you're if you're not confident, not sure because that gives you an idea. It's like a good starting point to build out uh what you need. Once you get one right, you also be more familiar with our platform as well. Don't get like so overwhelmed. Yeah. Okay. I think we're at the end today. Thank you everyone for joining. Uh I'm happy to answer all your questions. Look forward to the resources uh tomorrow and I hope more importantly more than anything that you guys get inspired today after joining us that you can build out something cool and interesting like even at your plan and even at your um you know your familiarity with AI it is possible. It's doable. Um, and like these are ways that you can like and like the best practices that you learned today should help you get further than you think you could uh to to create really powerful tools for uh your marketing efforts and analysis. Okay, thank you everyone for joining us today. Um, when you guys close um this this uh Zoom, you will be shown a page to fill out a survey to give feedback. Please uh share your feedback about today's session, what you liked about it, what you think we should improve. Uh you can also add in more information about like what you like to see in in future uh live sessions that we can cover. Um I've set the duration to be 30 minutes uh and then the rest of it like we're at an hour now, but that was like Q&A. So hopefully this was like ideal for most of you guys, but if you want that to be different, we can also accommodate that. Just let us know. Okay. Uh with that said, I will now go ahead and end today's webinar. I thank you guys for joining us. Look forward to the the the things tomorrow and future uh live sessions, workshops to come. We have more uh coming up in the the following weeks. Just look out for the announcements. Uh and I hope that you guys go um get started vibe coding. Improve two-hour webinar. Maybe we'll see if there's enough demand for two-hour webinar. If you guys really want to sit for two hours with me, I'll be happy to share my time. All right, guys. I'll head out now. I hope you guys have a great rest of the week. You take care and see you guys in the next