Q: Will more websites be supported going forward for review/comment scraping?
Or is the tool able to be customised for each individual site?
Am interested in using this to scrape forum posts for comments and run sentiment analysis on those comments. But each forum is structured differently with its own nuances.
A: Hi John! Let me explain why we need to go and add mediums one by one, first:
- Most of the scraper tools have the problem of getting content because the scraper can't know which texts to get in a page, as there are too many of them. That is why, we are focused to go one by one and add main stream channels first. So when we support a medium, the extension knows where the reviews are and how to get it. After that, we might dive into forums. So keep connecting with us with your comments 🤩
To help you with that, I might have another solution. We often try other scrapers and write articles on how to use them. Here is one of the scrapers we like: https://kimola.com/cognitive/resources/how-to-generate-an-nlp-dataset-from-any-internet-source
When you have an .xls file, you're ready to drag & drop it to Cognitive dashboard to analyze content.
Also, I'd love to know which forums they are. Can you shoot me an e-mail with the forums list? beybin@kimola.com we have a very curious team 😇
Thank you for submitting! To keep our community as helpful as possible, we moderate all comments and reviews.
Hi John! Let me explain why we need to go and add mediums one by one, first:
- Most of the scraper tools have the problem of getting content because the scraper can't know which texts to get in a page, as there are too many of them. That is why, we are focused to go one by one and add main stream channels first. So when we support a medium, the extension knows where the reviews are and how to get it. After that, we might dive into forums. So keep connecting with us with your comments 🤩
To help you with that, I might have another solution. We often try other scrapers and write articles on how to use them. Here is one of the scrapers we like: https://kimola.com/cognitive/resources/how-to-generate-an-nlp-dataset-from-any-internet-source
When you have an .xls file, you're ready to drag & drop it to Cognitive dashboard to analyze content.
Also, I'd love to know which forums they are. Can you shoot me an e-mail with the forums list? beybin@kimola.com we have a very curious team 😇
Hey Beybin thanks for such a helpful response. That is indeed an issue I've faced when using various web scraping tools for data extraction, and normally have to spend abit of time cleaning up the data before it is usable. Surprisingly enough, the Instant Data Scraper that you referred me to actually works perfectly for what I've been trying to do, with the least amount of cleaning required! Won't be much work to get the xls file ready to drop into the dashboard. This has very likely tipped me over the edge to grab a couple of codes and try out Kimola. Thanks again!
I've great news! We have increased our limits up to a MILLION! Starting from 100K, our limits are 10X higher now!
We love the Sumo-lings community, I hope this move shows our passion to become a lovemark! 🤩
Q: Why so overpriced?
Competitors charge more than twice as cheap and without any offer. Also, no renewal on queries per month, etc. It looks nice product to try, but not willing to commit to your pricing model.
Hi Tom! Our queries renew monthly :) But we're discussing to give higher queries for Sumo-lings. What would be the average query limit you'd like to see on packages?
I can relate to the pricing issue, but every product has its own features, for example if you're using one of the other products, you also have to pay to a scraper to get the data. With Kimola Cognitive, we are adding many mediums every day to grab data with our browser extension, for free! None of the ML platforms provide that.
And the other thing is Kimola has great pre-built ML models and new models are published regularly. So users don't have to create custom models to classify their data. See the Gallery here: https://kimola.com/cognitive/gallery
We see this deal as an entrance of our SaaS product family as we're building a data analytics company to reveal consumer insights. So we're not just offering a machine learning platform but a good start to integrate with our other products in the future that you can track your brand mentions, collect real-time customer feedbacks and analyze lifestyle of consumers and more. You can see kimola.com/analytics page for the first product of ours.
So we genuinely believe it's a great deal ❤️
I'm pretty sure they have great features too, but in the end we can not manage our pricing by just considering a LTD deal -what we can do is maybe to increase limits of users and we're considering that :)
Thank you for your help, hope we can have more discussions in the future :)
We have an API, you can take a look at the documentation here: https://github.com/Kimola/kimola-cognitive-api-documentation
We just met Pabbly, it looks like a great platform, we are discussing if we'd like to get it on the roadmap!
We are considering the higher limits but we don't have any use cases from users, would you be able to tell what would be the use case of yours if you had higher limits? We're trying to understand individual professional and that would mean so much to us :)
In terms of use cases, I work with SaaS orgs and having competitor data e.g. from Capterra or g2 would be very nice to guide how I work/consult with them (let's say I do a category comparison for them to get an overview - that can be several thousands of queries per client). In theory Level 2 should be just about enough (that is if I don't do anything else with the platform), but I think people are not really complaining about the limits, more like the cost of those limits... given that other and arguably more mature product was ~50% cheaper. Most likely most people just don't like paying 2x ... my thoughts, I could be wrong of course :)
Hi, I've great news! We have increased our limits up to a MILLION! Starting from 100K, our limits are 10X higher now!
We love the Sumo-lings community, I hope this move shows our passion to become a lovemark! 🤩
Q: Is there a bug or the product isn't working?
... I couldn't get anything to work, the chrome extension needs and API key which I don't have ... you actually can't get the reviews from Appsumo, the chrome extension can't see the reviews on AS .... and another thing, why do you guys mention G2? in the github readme file there is no such thing
A: Hi Daniel, you can find the API key in the "Models" menu on Cognitive dashboard. It's on the right corner at Models page (cognitive.kimola.com/models) and we're publishing a new read me file and a video to show how the browser extension works. You're so right to be frustrated and thank you for being patient :)
We support many mediums, some of them are changing their designs right now, just like Appsumo. We will be fixing the issue after their design is all changed.
Please let me know if you have any other issues, we're currently on live chat, giving support :) Also you can reach me out at beybin@kimola.com
Hi Daniel, you can find the API key in the "Models" menu on Cognitive dashboard. It's on the right corner at Models page (https://cognitive.kimola.com/models) and we're publishing a new read me file and a video to show how the browser extension works. You're so right to be frustrated and thank you for being patient :)
We support many mediums, some of them are changing their designs right now, just like Appsumo. We will be fixing the issue after their design is all changed.
Please let me know if you have any other issues, we're currently on live chat, giving support :) Also you can reach me out at beybin@kimola.com
Hi Daniel, as promised, Appsumo issue is fixed! Now you can grab comments from Appsumo whether it's the old or the new design. Just wanted to let you know :)
We keep growing with you :)
Bests.
Q: Hello,
I have a few questions before I purchase:
1.)
How do you measure queries?
For instance, if I went on Capterra and scraped 300 reviews about my product, would that count as 300 queries? Or do you count queries based on a specific amount of text characters? If so, how many?
1b.) When we use Kimola to retrieve the assembled data in the dashboard, or search that data, do those count as queries too? I like math, and I’m trying to get a sense of my upper limits on Tier 3.
For instance, if a query equals 50 characters and I have 100,000 queries, then I can estimate that to be 5 million characters. The average paragraph uses 300 characters, so that would be roughly 16,660 paragraphs. Divided by 3 for the average G2/Capterra review, and that would be 5,550 reviews.
Do you use a metric like this for queries?
2.) If we build a custom ML model, is that completely no-code? How does that process work? Do we use code blocks to assemble it? What happens if we can’t build what we want? Do you step in and build it for us? I imagine there are some limits to that.
3.) Do you have documentation for using all the elements of the platform? I noticed on your Youtube channel you’ve only uploaded 8 minutes of video in the last 2 months and none of it includes a voice. That is not useful at all when learning a new technology. Any plans for more elaborate training?
4.) As a use-case, do you think the following would be possible:
If I have 300 distinctly different products but some of them have similar features, could I input each product with feature keywords into Kimola, then search my products by keyword, based on what my customer is looking for, and get a list back of the best product matches?
5.) If I buy Tier 3, with 100,000 queries and I go over that limit, what happens? Can I buy more queries? If so, as an Appsumo customer would I receive a substantial discount on extra queries?
That’s it for now. Thank you so much for your answers and for coming to AS!! I was just looking online today for a tool like this.
A: Hi, I'm glad to tell that we started shooting the videos with audio :) Here is the Youtube playlist and more videos are on the way! --> https://bit.ly/Kimola-Cognitive-Walkthrough-Videos
Hi, I love it when a Sumo-ling asks sooo qualified questions, so thank you a lot! Please drop me an e-mail if you need more assistance: beybin@kimola.com
1.) Measuring queries A query is only spend when you analyze data. So scraping data or training AI doesn't spend from your queries. Each analyzed row is one query and each row should be 1000 characters at most. The system tells how many queries will be spent before each analysis.
1b.) When we use Kimola to retrieve the assembled data in the dashboard, or search that data, do those count as queries too? I like math, and I’m trying to get a sense of my upper limits on Tier 3. You can only spend queries by analyzing texts.
2.) If we build a custom ML model, is that completely no-code? Completely no code. I created tens of models before without coding :) All you need to do is drag and drop your .xls file and the AI will be trained by looking at the content & label. Afterwards, when you upload your raw .xls file that contains feedbacks to be analyzed, it creates the report on the dashboard for you to view and jiggle with it. You can see the GIF while I'm creating and training a custom model here: https://kimola.com/support/cognitive/how-to-create-a-custom-machine-learning-model I'm also shooting a video for Sumo-lings and I'll upload it to Youtube later today for this issue and a couple of more. We also give support if you need any help, so don't worry, you won't be alone :) We are a very responsive team.
3.) Do you have documentation for using all the elements of the platform? I noticed on your Youtube channel you’ve only uploaded 8 minutes of video in the last 2 months and none of it includes a voice. That is not useful at all when learning a new technology. Any plans for more elaborate training? - We will be shooting 15 "Do it yourself" videos later this month with a big production. - We're also shooting quick videos today for Sumo-lings :) - We have great step by step support articles with GIF's: https://kimola.com/cognitive/support
4.) As a use-case, do you think the following would be possible: If I have 300 distinctly different products but some of them have similar features, could I input each product with feature keywords into Kimola, then search my products by keyword, based on what my customer is looking for, and get a list back of the best product matches?
When you upload and analyze a dataset, you can take advantage of the search feature -search doesn't spend from queries. There is a powerful system that supports wildcard search and updates the analysis results live according to the search results.
5.) If I buy Tier 3, with 100,000 queries and I go over that limit, what happens? Can I buy more queries? If so, as an Appsumo customer would I receive a substantial discount on extra queries? We notify with e-mail before you reach out to your limits. Afterwards, additional 10K queries are added to the package for $49 for tier 3.
We'd be so glad to have you on Kimola Cognitive :)
Q: "It extracts reviews on supported mediums such as Youtube, Capterra, Tripadvisor etc."
Kimola Airset Generator currently supports; Youtube, Capterra, AppSumo, AirBnB, Reddit
So it doesn't support tripadvisor.
Overall it looks it's more for saas and travel related field.
Tried to review videos you guys have on your youtube channel. They don't have any audio.
If you could guys explain a bit more the benefits that would be great.
A: Hi, browser extension is only one of the features of our no-code machine learning platform. If you already have a dataset as an .xls file, you can analyze any text data by simply dragging and dropping to the dashboard 😇
We have clients use our solution in different areas such as FMCG, Sports, Retail, Pharma.
Our videos are connected t, our support articles but we will be adding more with audio <3 You can view all of our articles in here: https://kimola.com/support/cognitive
I think one of my favorite things is our Cognitive Gallery: You can analyze text data around the models that we have on the gallery, which are SaaS, Mobile Apps, Mobile Games, Common Consumer Conversations Classifier (Enterprise brands use this one), Banking, Telco, Automative. We regularly publish pre-built models to help our users analyze their customer feedbacks better.
You can also create your own custom models too, with a simple drag & drop.
If you'd like to learn more, we can make a walkthrough 😇
Hi, I'm glad to tell that we started shooting the videos! Here is the Youtube playlist and more videos are on the way! --> https://bit.ly/Kimola-Cognitive-Walkthrough-Videos
Q: I definitely like this product!
What would make me instantly buy this is if you combined the Analytics and Cognitive tools. Providing access to both of these would be a “GOLDEN TACO” deal. Is this possible?
Please make a tutorial video with audio. The ones that I have seen on YouTube aren’t useful (to me) without audio. Are you able to do this?
Also, the Tier 1 package needs more punch🥊🥊🥊🥊🥊. Is it possible for you to make this more attractive?
A: Hi, I'm glad to tell that we started shooting the videos! Here is the Youtube playlist and more videos are on the way! --> https://bit.ly/Kimola-Cognitive-Walkthrough-Videos
Soo glad to get this comment! We would love Sumolings to try Analytics! Unfortunately we can’t bundle a deal right now as we’re already live but we would love to give a surprise if you reach out: support@kimola.com
I’m going to shoot a video tomorrow as there are too many questions on live chat today, hope everyone likes my voice 🎤
I’m not sure if we can add new things after the deal is live but I’m going to ask this to Appsumo team and talk with Kimola team right away! Will write back soon ^^
So the row / character calculations are referred to query count. If you have more than 1000 characters in a row, it still analyzes it but you'd spend a second query after 1000 characters.
This seems like a very interesting purchase for me, and I have no problems with the limits and so on, because I understand the costs of a quality service like this, but what worries me is the language. My target audience and clients essentially speak Spanish. Is there a roadmap or something to see, and more specifically, is there a possibility that the Spanish language will be added in the short term? Thanks so much for reading.
A: Thank you for the positive vibes, we’re really excited to have you! Spanish is already one of our supported languages ❤️
Also, of course we have a roadmap and we would love Sumo-lings to like or comment on the upcoming features and ideas on our roadmap: https://bit.ly/kimola-roadmap.
But then you would have to spend some time on cleaning so it would be hard :))) Also, i think there are tools like Semrush who does that for websites, they have an SEO angle which might be so much better.
Q: Can I upload a pdf of a book or business proposal and get an analysis like it were 1000 characters per query?
so if the pdf is 1M characters for it to consumer 1000 credits.
Loved the idea but not sure about the insights to be revealed, please let me know if you've a text data like this in .xls file and we can create a report to see the results!
Let's say you want to see what content to make or what language your competitors are using. a wordcloud will help you understand the jargon of a particular business from their domain name. I guess that's what cold outreach AI uses to generate proposals, but it's the same tech you use.
The insight would be helpful to know what to talk about when you outreach to a website. I understand the use cases can be limited, though
I can see how useful it can be now but there are too many words on a website to make it SEO-wise, so it would might be a mess.
Rather than that, I think I'd go to their Twitter account and scrape their tweets! This way I can also understand their communications strategy as an extra! What would you think about that?
I understand not the whole robots.txt of a website, but just rather a particular blog post or page, very few pages are that long.
So if targeting examplewebsite.com/home , it doesn't have to crawl examplewebsite.com/team too
Q: Can this be used to understand websites?
So like I have a list of domains, and I want the tool to scrape the text and generate a wordcloud for me to understand the client well very quickly and in bulk
Let me explain why we need to go and add mediums one by one, first:
- Most of the scraper tools have the problem of getting content because the scraper can't know which texts to get in a page, as there are too many of them. That is why, we are focused to go one by one and add main stream channels first. So when we support a medium, the extension knows where the reviews are and how to get it.
To help you with that, I might have another solution. We often try other scrapers and write articles on how to use them. Here is one of the scrapers we like: https://kimola.com/cognitive/resources/how-to-generate-an-nlp-dataset-from-any-internet-source
When you have an .xls file, you're ready to drag & drop it to Cognitive dashboard to analyze content.
Yeah, but I want it to scrape the text from a website for me, for example from a website. https://www.wordclouds.com/ does it for free from a domain but I want to be able to do this in bulk. It is annoying to go one by one.
A: Sure, you can comment and like on Github from it's own page. Follow these steps: Visit here first: https://github.com/orgs/Kimola/projects/4 Click on one of the ideas that you'd like to comment A pop-up will appear :)
Of course we have a roadmap and we would love Sumo-lings to like or comment on the upcoming features and ideas on our roadmap: https://bit.ly/kimola-roadmap.
Deeptalk and its team is great, but we have a lot to offer too. 💖
As soon as you have access to Kimola Cognitive, you can start using pre-built, ready-to-use models on various topics created by the Kimola team. To explore the Gallery, you can go with this link: https://kimola.com/cognitive/gallery
In addition, Kimola Cognitive provides its users with an extension from which they can grab data from various sites. https://kimola.com/support/cognitive/setting-up-the-airset-browser-extension
Also, you can classify your data sets with multiple ready-to-use ML models at the same time and benefit from text analysis.
A: Hi, thanks for the question! Such a great community here ✨
Kimola Cognitive is designed to analyze all kinds of text-based data. Kimola Cognitive has a term-extraction technology based word-cloud feature (it can be found on the reports page.) It's not just a simple word cloud but the system creates the word cloud and generates a filter from the popular topics to view content.
Yes, you can do sentiment analysis with your text data. There is not only a ready-to-use, pre-built sentiment ML model but also many other models in our gallery (kimola.com/cognitive/gallery) that you can just drag and drop your text data and start the analysis in 3 minutes. You can use these sector specific (Horeca, SaaS, App Review Models) and general models (Sentiment, Hate Speech Detection and News Classification, etc.) to analyze your data too. See how to do it on this article: https://kimola.com/support/cognitive/using-kimola-cognitive-gallery-and-pre-built-ml-models
You can train your custom ML model with your labeled data with a simple drag & drop. For further more, you can read this article: https://kimola.com/support/cognitive/how-to-create-a-custom-machine-learning-model
Unfortunately, Kimola Cognitive only supports XLSX, XLS, TSY, and CSV file formats.
Airset is currently supports YouTube, Capterra, AppSumo, AirBnB and Reddit. You can read mode from here about Airset: https://github.com/Kimola/airset-generator#readme
You can read the article below to understand what is model and how does it work. Article: https://kimola.com/support/cognitive/what-is-a-machine-learning-model
Kimola Cognitive supports 6 languages right now; English, Spanish, Portuguese, French and German. We update the list in support when we add a new one: https://kimola.com/support/cognitive/supported-languages
Indonesian might be interesting though! ıt's not on our road map, but will consider it after seeing the requests here ;)
Supported mediums are on here: https://github.com/Kimola/airset-generator#readme
Browser extension is not able to scrape Facebook but Instagram and Reddit. On Instagram, you can scrape the comments of a post. On Reddit, you can scrape comments of a post.
Q: Will more websites be supported going forward for review/comment scraping?
Or is the tool able to be customised for each individual site?
Am interested in using this to scrape forum posts for comments and run sentiment analysis on those comments. But each forum is structured differently with its own nuances.
Beybin_KimolaCognitive
May 15, 2024A: Hi John! Let me explain why we need to go and add mediums one by one, first:
- Most of the scraper tools have the problem of getting content because the scraper can't know which texts to get in a page, as there are too many of them. That is why, we are focused to go one by one and add main stream channels first. So when we support a medium, the extension knows where the reviews are and how to get it. After that, we might dive into forums. So keep connecting with us with your comments 🤩
To help you with that, I might have another solution. We often try other scrapers and write articles on how to use them. Here is one of the scrapers we like: https://kimola.com/cognitive/resources/how-to-generate-an-nlp-dataset-from-any-internet-source
When you have an .xls file, you're ready to drag & drop it to Cognitive dashboard to analyze content.
Also, I'd love to know which forums they are. Can you shoot me an e-mail with the forums list? beybin@kimola.com we have a very curious team 😇
Thank you for submitting! To keep our community as helpful as possible, we moderate all comments and reviews.
Share Kimola Cognitive
Hi John! Let me explain why we need to go and add mediums one by one, first:
- Most of the scraper tools have the problem of getting content because the scraper can't know which texts to get in a page, as there are too many of them. That is why, we are focused to go one by one and add main stream channels first. So when we support a medium, the extension knows where the reviews are and how to get it. After that, we might dive into forums. So keep connecting with us with your comments 🤩
To help you with that, I might have another solution. We often try other scrapers and write articles on how to use them. Here is one of the scrapers we like: https://kimola.com/cognitive/resources/how-to-generate-an-nlp-dataset-from-any-internet-source
When you have an .xls file, you're ready to drag & drop it to Cognitive dashboard to analyze content.
Also, I'd love to know which forums they are. Can you shoot me an e-mail with the forums list? beybin@kimola.com we have a very curious team 😇
Hey Beybin thanks for such a helpful response. That is indeed an issue I've faced when using various web scraping tools for data extraction, and normally have to spend abit of time cleaning up the data before it is usable. Surprisingly enough, the Instant Data Scraper that you referred me to actually works perfectly for what I've been trying to do, with the least amount of cleaning required! Won't be much work to get the xls file ready to drop into the dashboard. This has very likely tipped me over the edge to grab a couple of codes and try out Kimola. Thanks again!
John, thank you for the positive vibes :) I'm looking forward to hear your honest reviews :)
Let me know if you need anything else, here is me: beybin@kimola.com
Hi John,
I've great news! We have increased our limits up to a MILLION! Starting from 100K, our limits are 10X higher now!
We love the Sumo-lings community, I hope this move shows our passion to become a lovemark! 🤩
Q: Why so overpriced?
Competitors charge more than twice as cheap and without any offer. Also, no renewal on queries per month, etc. It looks nice product to try, but not willing to commit to your pricing model.
Beybin_KimolaCognitive
May 15, 2024A: Hi, I've great news! We have increased our limits up to a MILLION! Starting from 100K, our limits are 10X higher now!
We love the Sumo-lings community, I hope this move shows our passion to become a lovemark! 🤩
Share Kimola Cognitive
Hi Tom! Our queries renew monthly :) But we're discussing to give higher queries for Sumo-lings. What would be the average query limit you'd like to see on packages?
Verified purchaser
Similar to the one offered with Deep-talk
Good to know! Let us consider this with Appsumo team & Kimola team :)
Q: Hello, looks very interesting thanks for coming to AS!
:)
Is it possible to send data to other applications or websites? API? Integrations (e.g. Pabbly)?
And after reading other comments, also have to agree that higher limits would be nice! :)
Thanks!
Beybin_KimolaCognitive
May 15, 2024A: Great to learn those cases, thanks!
I can relate to the pricing issue, but every product has its own features, for example if you're using one of the other products, you also have to pay to a scraper to get the data. With Kimola Cognitive, we are adding many mediums every day to grab data with our browser extension, for free! None of the ML platforms provide that.
And the other thing is Kimola has great pre-built ML models and new models are published regularly. So users don't have to create custom models to classify their data. See the Gallery here: https://kimola.com/cognitive/gallery
We see this deal as an entrance of our SaaS product family as we're building a data analytics company to reveal consumer insights. So we're not just offering a machine learning platform but a good start to integrate with our other products in the future that you can track your brand mentions, collect real-time customer feedbacks and analyze lifestyle of consumers and more. You can see kimola.com/analytics page for the first product of ours.
So we genuinely believe it's a great deal ❤️
I'm pretty sure they have great features too, but in the end we can not manage our pricing by just considering a LTD deal -what we can do is maybe to increase limits of users and we're considering that :)
Thank you for your help, hope we can have more discussions in the future :)
Share Kimola Cognitive
Hi, thank you for the positive vibes :)
We have an API, you can take a look at the documentation here: https://github.com/Kimola/kimola-cognitive-api-documentation
We just met Pabbly, it looks like a great platform, we are discussing if we'd like to get it on the roadmap!
We are considering the higher limits but we don't have any use cases from users, would you be able to tell what would be the use case of yours if you had higher limits? We're trying to understand individual professional and that would mean so much to us :)
Thanks!
Verified purchaser
Good! :))
In terms of use cases, I work with SaaS orgs and having competitor data e.g. from Capterra or g2 would be very nice to guide how I work/consult with them (let's say I do a category comparison for them to get an overview - that can be several thousands of queries per client). In theory Level 2 should be just about enough (that is if I don't do anything else with the platform), but I think people are not really complaining about the limits, more like the cost of those limits... given that other and arguably more mature product was ~50% cheaper. Most likely most people just don't like paying 2x ... my thoughts, I could be wrong of course :)
Hi, I've great news! We have increased our limits up to a MILLION! Starting from 100K, our limits are 10X higher now!
We love the Sumo-lings community, I hope this move shows our passion to become a lovemark! 🤩
Q: Is there a bug or the product isn't working?
... I couldn't get anything to work, the chrome extension needs and API key which I don't have ... you actually can't get the reviews from Appsumo, the chrome extension can't see the reviews on AS .... and another thing, why do you guys mention G2? in the github readme file there is no such thing
Beybin_KimolaCognitive
May 15, 2024A: Hi Daniel, you can find the API key in the "Models" menu on Cognitive dashboard. It's on the right corner at Models page (cognitive.kimola.com/models) and we're publishing a new read me file and a video to show how the browser extension works. You're so right to be frustrated and thank you for being patient :)
We support many mediums, some of them are changing their designs right now, just like Appsumo. We will be fixing the issue after their design is all changed.
Please let me know if you have any other issues, we're currently on live chat, giving support :) Also you can reach me out at beybin@kimola.com
Share Kimola Cognitive
Hi Daniel, you can find the API key in the "Models" menu on Cognitive dashboard. It's on the right corner at Models page (https://cognitive.kimola.com/models) and we're publishing a new read me file and a video to show how the browser extension works. You're so right to be frustrated and thank you for being patient :)
We support many mediums, some of them are changing their designs right now, just like Appsumo. We will be fixing the issue after their design is all changed.
Please let me know if you have any other issues, we're currently on live chat, giving support :) Also you can reach me out at beybin@kimola.com
Hi Daniel, as promised, Appsumo issue is fixed! Now you can grab comments from Appsumo whether it's the old or the new design. Just wanted to let you know :)
We keep growing with you :)
Bests.
Q: Hello, I have a few questions before I purchase: 1.)
How do you measure queries?
For instance, if I went on Capterra and scraped 300 reviews about my product, would that count as 300 queries? Or do you count queries based on a specific amount of text characters? If so, how many?
1b.) When we use Kimola to retrieve the assembled data in the dashboard, or search that data, do those count as queries too? I like math, and I’m trying to get a sense of my upper limits on Tier 3.
For instance, if a query equals 50 characters and I have 100,000 queries, then I can estimate that to be 5 million characters. The average paragraph uses 300 characters, so that would be roughly 16,660 paragraphs. Divided by 3 for the average G2/Capterra review, and that would be 5,550 reviews.
Do you use a metric like this for queries?
2.) If we build a custom ML model, is that completely no-code? How does that process work? Do we use code blocks to assemble it? What happens if we can’t build what we want? Do you step in and build it for us? I imagine there are some limits to that.
3.) Do you have documentation for using all the elements of the platform? I noticed on your Youtube channel you’ve only uploaded 8 minutes of video in the last 2 months and none of it includes a voice. That is not useful at all when learning a new technology. Any plans for more elaborate training?
4.) As a use-case, do you think the following would be possible:
If I have 300 distinctly different products but some of them have similar features, could I input each product with feature keywords into Kimola, then search my products by keyword, based on what my customer is looking for, and get a list back of the best product matches?
5.) If I buy Tier 3, with 100,000 queries and I go over that limit, what happens? Can I buy more queries? If so, as an Appsumo customer would I receive a substantial discount on extra queries?
That’s it for now. Thank you so much for your answers and for coming to AS!! I was just looking online today for a tool like this.
Beybin_KimolaCognitive
May 15, 2024A: Hi, I'm glad to tell that we started shooting the videos with audio :) Here is the Youtube playlist and more videos are on the way! --> https://bit.ly/Kimola-Cognitive-Walkthrough-Videos
Share Kimola Cognitive
Hi, I love it when a Sumo-ling asks sooo qualified questions, so thank you a lot! Please drop me an e-mail if you need more assistance: beybin@kimola.com
1.) Measuring queries
A query is only spend when you analyze data. So scraping data or training AI doesn't spend from your queries. Each analyzed row is one query and each row should be 1000 characters at most. The system tells how many queries will be spent before each analysis.
1b.) When we use Kimola to retrieve the assembled data in the dashboard, or search that data, do those count as queries too? I like math, and I’m trying to get a sense of my upper limits on Tier 3.
You can only spend queries by analyzing texts.
2.) If we build a custom ML model, is that completely no-code?
Completely no code. I created tens of models before without coding :) All you need to do is drag and drop your .xls file and the AI will be trained by looking at the content & label. Afterwards, when you upload your raw .xls file that contains feedbacks to be analyzed, it creates the report on the dashboard for you to view and jiggle with it. You can see the GIF while I'm creating and training a custom model here: https://kimola.com/support/cognitive/how-to-create-a-custom-machine-learning-model
I'm also shooting a video for Sumo-lings and I'll upload it to Youtube later today for this issue and a couple of more. We also give support if you need any help, so don't worry, you won't be alone :) We are a very responsive team.
3.) Do you have documentation for using all the elements of the platform? I noticed on your Youtube channel you’ve only uploaded 8 minutes of video in the last 2 months and none of it includes a voice. That is not useful at all when learning a new technology. Any plans for more elaborate training?
- We will be shooting 15 "Do it yourself" videos later this month with a big production.
- We're also shooting quick videos today for Sumo-lings :)
- We have great step by step support articles with GIF's: https://kimola.com/cognitive/support
4.) As a use-case, do you think the following would be possible:
If I have 300 distinctly different products but some of them have similar features, could I input each product with feature keywords into Kimola, then search my products by keyword, based on what my customer is looking for, and get a list back of the best product matches?
When you upload and analyze a dataset, you can take advantage of the search feature -search doesn't spend from queries. There is a powerful system that supports wildcard search and updates the analysis results live according to the search results.
5.) If I buy Tier 3, with 100,000 queries and I go over that limit, what happens? Can I buy more queries? If so, as an Appsumo customer would I receive a substantial discount on extra queries?
We notify with e-mail before you reach out to your limits. Afterwards, additional 10K queries are added to the package for $49 for tier 3.
We'd be so glad to have you on Kimola Cognitive :)
Q: "It extracts reviews on supported mediums such as Youtube, Capterra, Tripadvisor etc."
Kimola Airset Generator currently supports;
Youtube, Capterra, AppSumo, AirBnB, Reddit
So it doesn't support tripadvisor.
Overall it looks it's more for saas and travel related field.
Tried to review videos you guys have on your youtube channel. They don't have any audio.
If you could guys explain a bit more the benefits that would be great.
Beybin_KimolaCognitive
May 15, 2024A: Hi, browser extension is only one of the features of our no-code machine learning platform. If you already have a dataset as an .xls file, you can analyze any text data by simply dragging and dropping to the dashboard 😇
We have clients use our solution in different areas such as FMCG, Sports, Retail, Pharma.
Our videos are connected t, our support articles but we will be adding more with audio <3 You can view all of our articles in here: https://kimola.com/support/cognitive
I think one of my favorite things is our Cognitive Gallery: You can analyze text data around the models that we have on the gallery, which are SaaS, Mobile Apps, Mobile Games, Common Consumer Conversations Classifier (Enterprise brands use this one), Banking, Telco, Automative. We regularly publish pre-built models to help our users analyze their customer feedbacks better.
You can also create your own custom models too, with a simple drag & drop.
If you'd like to learn more, we can make a walkthrough 😇
Share Kimola Cognitive
Hi, I'm glad to tell that we started shooting the videos! Here is the Youtube playlist and more videos are on the way! --> https://bit.ly/Kimola-Cognitive-Walkthrough-Videos
Q: I definitely like this product!
What would make me instantly buy this is if you combined the Analytics and Cognitive tools. Providing access to both of these would be a “GOLDEN TACO” deal. Is this possible?
Please make a tutorial video with audio. The ones that I have seen on YouTube aren’t useful (to me) without audio. Are you able to do this?
Also, the Tier 1 package needs more punch🥊🥊🥊🥊🥊. Is it possible for you to make this more attractive?
Beybin_KimolaCognitive
May 15, 2024A: Hi, I'm glad to tell that we started shooting the videos! Here is the Youtube playlist and more videos are on the way! --> https://bit.ly/Kimola-Cognitive-Walkthrough-Videos
Share Kimola Cognitive
Soo glad to get this comment! We would love Sumolings to try Analytics! Unfortunately we can’t bundle a deal right now as we’re already live but we would love to give a surprise if you reach out: support@kimola.com
I’m going to shoot a video tomorrow as there are too many questions on live chat today, hope everyone likes my voice 🎤
I’m not sure if we can add new things after the deal is live but I’m going to ask this to Appsumo team and talk with Kimola team right away! Will write back soon ^^
😉
Q: is it possible to increase the amount of row size?
so more than 1000 characters per row.
Share Kimola Cognitive
Verified purchaser
My use case is to upload call transcripts of company earnings etc
I thought I replied this, sorry Daniel!
So the row / character calculations are referred to query count. If you have more than 1000 characters in a row, it still analyzes it but you'd spend a second query after 1000 characters.
Verified purchaser
makes sense, thanks!
Q: Hello beautiful people of Kimola!!!
This seems like a very interesting purchase for me, and I have no problems with the limits and so on, because I understand the costs of a quality service like this, but what worries me is the language.
My target audience and clients essentially speak Spanish.
Is there a roadmap or something to see, and more specifically, is there a possibility that the Spanish language will be added in the short term?
Thanks so much for reading.
Beybin_KimolaCognitive
May 15, 2024A: Thank you for the positive vibes, we’re really excited to have you!
Spanish is already one of our supported languages ❤️
Also, of course we have a roadmap and we would love Sumo-lings to like or comment on the upcoming features and ideas on our roadmap: https://bit.ly/kimola-roadmap.
Share Kimola Cognitive
Q: Does scraping the reviews count as a query, then another query analyzing them?
or how does that work.
Beybin_KimolaCognitive
May 15, 2024A: Nope, only analyzing texts spend query when you drag & drop or scrape the data to analyze.
Share Kimola Cognitive
But then you would have to spend some time on cleaning so it would be hard :))) Also, i think there are tools like Semrush who does that for websites, they have an SEO angle which might be so much better.
Q: Can I upload a pdf of a book or business proposal and get an analysis like it were 1000 characters per query?
so if the pdf is 1M characters for it to consumer 1000 credits.
Could you build this feature?¿
Beybin_KimolaCognitive
May 14, 2024A: Hi Dani, great question again :)
Here are the file formats that we support: XLSX, XLS, TSY, and CSV. So it should be one of these to drag & drop to analyze.
Share Kimola Cognitive
Verified purchaser
Yeah, you could build a feature to import larger text documents too though, could be interesting, could think about workarounds in the meantime
Loved the idea but not sure about the insights to be revealed, please let me know if you've a text data like this in .xls file and we can create a report to see the results!
Verified purchaser
Let's say you want to see what content to make or what language your competitors are using. a wordcloud will help you understand the jargon of a particular business from their domain name. I guess that's what cold outreach AI uses to generate proposals, but it's the same tech you use.
The insight would be helpful to know what to talk about when you outreach to a website. I understand the use cases can be limited, though
I can see how useful it can be now but there are too many words on a website to make it SEO-wise, so it would might be a mess.
Rather than that, I think I'd go to their Twitter account and scrape their tweets! This way I can also understand their communications strategy as an extra! What would you think about that?
Verified purchaser
Interesting, yes and it makes sense.
I understand not the whole robots.txt of a website, but just rather a particular blog post or page, very few pages are that long.
So if targeting examplewebsite.com/home , it doesn't have to crawl examplewebsite.com/team too
Q: Can this be used to understand websites?
So like I have a list of domains, and I want the tool to scrape the text and generate a wordcloud for me to understand the client well very quickly and in bulk
Beybin_KimolaCognitive
May 15, 2024A: Ha Daniel,
Let me explain why we need to go and add mediums one by one, first:
- Most of the scraper tools have the problem of getting content because the scraper can't know which texts to get in a page, as there are too many of them. That is why, we are focused to go one by one and add main stream channels first. So when we support a medium, the extension knows where the reviews are and how to get it.
To help you with that, I might have another solution. We often try other scrapers and write articles on how to use them. Here is one of the scrapers we like: https://kimola.com/cognitive/resources/how-to-generate-an-nlp-dataset-from-any-internet-source
When you have an .xls file, you're ready to drag & drop it to Cognitive dashboard to analyze content.
Hope this helps ❤️
Share Kimola Cognitive
Hi, if you have the texts in an .xls file, yes, you can reveal insights but extension is only working for specific mediums.
Verified purchaser
Yeah, but I want it to scrape the text from a website for me, for example from a website. https://www.wordclouds.com/ does it for free from a domain but I want to be able to do this in bulk. It is annoying to go one by one.
Q: Spanish would me an amazing language to have.
Beybin_KimolaCognitive
May 14, 2024A: Hi again 👋🏼
Kimola Cognitive supports 6 languages listed below and Spanish is one of them! 🍀
English
Spanish
French
German
Portuguese
Turkish
Please feel free to ask any question!
Share Kimola Cognitive
Verified purchaser
Amazing!
Q: Hi, where can I propose a new language to support?
Is there a roadmap?
Thank you
Beybin_KimolaCognitive
May 15, 2024A: Sure, you can comment and like on Github from it's own page. Follow these steps:
Visit here first: https://github.com/orgs/Kimola/projects/4
Click on one of the ideas that you'd like to comment
A pop-up will appear :)
Share Kimola Cognitive
Hi 👋🏼
Thank you so much for you interest!
Of course we have a roadmap and we would love Sumo-lings to like or comment on the upcoming features and ideas on our roadmap: https://bit.ly/kimola-roadmap.
Verified purchaser
IS it possible to upvote some of those ideas? Sorry if I make a dumb question but I am not confident with using Github
Thanks
Q: how do you compare to deep-talk.
ai?¿
Beybin_KimolaCognitive
May 15, 2024A: Hi again 🤗 Thank you for this question!
Deeptalk and its team is great, but we have a lot to offer too. 💖
As soon as you have access to Kimola Cognitive, you can start using pre-built, ready-to-use models on various topics created by the Kimola team. To explore the Gallery, you can go with this link: https://kimola.com/cognitive/gallery
In addition, Kimola Cognitive provides its users with an extension from which they can grab data from various sites.
https://kimola.com/support/cognitive/setting-up-the-airset-browser-extension
Also, you can classify your data sets with multiple ready-to-use ML models at the same time and benefit from text analysis.
Share Kimola Cognitive
Q: is this specifically for reviews or it can be used for other text based analysis?
¿ can you do wordcloud, sentiments, classify, negativity,....train your own models? upload a big pdf file and get the most used words and sentiment?¿
Beybin_KimolaCognitive
May 15, 2024A: Hi, thanks for the question! Such a great community here ✨
Kimola Cognitive is designed to analyze all kinds of text-based data. Kimola Cognitive has a term-extraction technology based word-cloud feature (it can be found on the reports page.) It's not just a simple word cloud but the system creates the word cloud and generates a filter from the popular topics to view content.
Yes, you can do sentiment analysis with your text data. There is not only a ready-to-use, pre-built sentiment ML model but also many other models in our gallery (kimola.com/cognitive/gallery) that you can just drag and drop your text data and start the analysis in 3 minutes. You can use these sector specific (Horeca, SaaS, App Review Models) and general models (Sentiment, Hate Speech Detection and News Classification, etc.) to analyze your data too. See how to do it on this article: https://kimola.com/support/cognitive/using-kimola-cognitive-gallery-and-pre-built-ml-models
You can train your custom ML model with your labeled data with a simple drag & drop. For further more, you can read this article: https://kimola.com/support/cognitive/how-to-create-a-custom-machine-learning-model
Unfortunately, Kimola Cognitive only supports XLSX, XLS, TSY, and CSV file formats.
Share Kimola Cognitive
Q: Can your app scrape Google reviews or Trustpilot reviews?
Also, what is a model and how does it work please?
Thank you!
Beybin_KimolaCognitive
May 15, 2024A: Hi, thanks for great question!
Airset is currently supports YouTube, Capterra, AppSumo, AirBnB and Reddit.
You can read mode from here about Airset: https://github.com/Kimola/airset-generator#readme
You can read the article below to understand what is model and how does it work.
Article: https://kimola.com/support/cognitive/what-is-a-machine-learning-model
Share Kimola Cognitive
Q: your Analytics product looks way more interesting, any chance this is offered on AppSumo?
Beybin_KimolaCognitive
May 15, 2024A: Mike, thanks, we love it too 🤩 We really loved the Appsumo team and the process. If we can do this right, we might consider a second campaign 🧚♀️
Share Kimola Cognitive
Hey mike.rosales.dpo,
This sounds like an interesting request, I am sure the partner will take this into consideration!
Q: Can you analyze multilingual conversations: English and Indonesian words?
Beybin_KimolaCognitive
May 15, 2024A: Hi,
Kimola Cognitive supports 6 languages right now; English, Spanish, Portuguese, French and German. We update the list in support when we add a new one: https://kimola.com/support/cognitive/supported-languages
Indonesian might be interesting though! ıt's not on our road map, but will consider it after seeing the requests here ;)
Share Kimola Cognitive
Q: Can it scrape facebook / instagram.
/ reddit comments to an excel file and also import into kimola?
Beybin_KimolaCognitive
May 15, 2024A: Hi!
Supported mediums are on here: https://github.com/Kimola/airset-generator#readme
Browser extension is not able to scrape Facebook but Instagram and Reddit.
On Instagram, you can scrape the comments of a post.
On Reddit, you can scrape comments of a post.
Share Kimola Cognitive