mxchat

Hi @inioluwaolaposi,

Thanks for reaching out. I’ve made a YouTube tutorial going over how to setup the Pinecone DB manager.

Pinecone is better because it’s designed for vector embeddings and cosine matching, making it much more efficient for large datasets. The default WordPress SQL database isn’t built for this, so we have to batch requests to 500 at a time to avoid crashing the site, which slows things down significantly. Pinecone handles large data seamlessly without this limitation, keeping your site fast and stable.

Let me know if you want me to explain any specific part in more detail or if you have any troubles.

Thanks again,

Maxwell

Thank you @tomchen29 for taking time out of your day to leave this kind review. I appreciate any and all feedback. It was thanks to you I accelerated the Pinecone integration and implemented a solution to handle larger data sets. Thank you again.

Maxwell

@merlynlabs Thank you so much for taking time of your day to leave this wonderful review. It very motivating and encouraging to see people enjoying the product, especially coming from someone with your experience. Thank you again!

Maxwell

?Muchísimas gracias por tomarte el tiempo de escribir esta rese?a tan maravillosa! Tu opinión y retroalimentación son muy valiosas para nosotros, y las apreciamos de corazón. Nos alegra saber que MxChat está siendo útil para ti. ?Gracias nuevamente por tu apoyo y recomendación!

Maxwell

Thanks for the kind words @deltellabogados! I’m beyond thankful you challenged me over a year ago to venture down this path. I have many updates planned and will continue to deliver the absolute best product and customer service possible.

Maxwell

Thank you so much for the review @bluepixel33! I’m glad you are enjoying the plugin. My only mission at the moment is to build the absolute best ai chatbot available for WordPress and maintain excellent customer service along the way. Your feedback is much appreciated, thank you.

Maxwell

Alrighty @tomchen29 the brand new MxChat 2.0 was just released! This includes a direction integration to Pinecone which solves two of your issues – scalability and bulk group deletion. Once you update to version 2.0 you will see a new add ons page which will show you how to download pinecone integration add on if you’d like. It’s completely free. It also includes a completely redesigned UI and branding that I’ve been working on.

Please let me know if you have any questions or troubles.

Thanks again,

Maxwell

@tomchen29 – This update will be released today, likely around 9PM US EST. It’s a major update setting the foundation for many other features and integrations to come.

If you are on my mailing list I will send an email out when the update goes live, explaining MxChat 2.0, and the future to come. Below are some sneak peek screenshots. This will be released to all users for free. Many more integrations coming for pro users.

Hi again @tomchen29,

Dropping in to let you know that I think you’re going to be blown away by this Pinecone integration. Can’t wait to show it to you. Super easy setup and takes less than a few minutes. It also has a very good database management UI in the WordPress admin panel. Very likely this could be ready by tomorrow.

Thanks,

Maxwell

• This reply was modified 1 month, 2 weeks ago by mxchat.

Hi @tomchen29,

I’m sure this is something that can be done. I’ll review it over the next few days. Also, Pinecone integration is coming very soon. I’ve already successfully integrated it on a test website. I may be able to push them both of those features in an update within a few days. If some issues come up with Pinecone, I’ll see about pushing this one faster.

Thanks,

Maxwell

• This reply was modified 1 month, 2 weeks ago by mxchat.

glad to hear it worked! I’ve been meaning to tackle the Pinecone integration as it will provide a much better performance for larger datasets. This is something I will start to work on over the next week. I don’t believe it will be very difficult so it shouldn’t take long.

If you’re enjoying MxChat, I’d really appreciate it if you could leave a review! Your feedback helps the project grow and lets others know what to expect. I take pride in delivering exceptional service and a quality product, and your support means a lot!

Thanks again for bringing this to my attention and helping MxChat become even better!

Maxwell

Morning @tomchen29,

I was able to replicate the same issue on my end this morning when pushing into the same amount of knowledge submissions. I tested my fix I had in mind and it did work for me. I went ahead and pushed the update this morning to 1.6.5. Please update to this and make sure you clear any cache just in case and test. If you update to 1.6.5 and it’s still not working please let me know so I can review any errors that I can see on my end.

Once we confirm it did work for you I’ll close this ticket (or you can) as resolved.

Over the next few weeks I will also explore implement an external database option such as Pinecone. As some users are wanting to push this to the limits and realistically if you want to have thousands and thousands of submission the most optimized and best approach is to use a database specifically used for Vector embeddings, such as Pinecone.

This solution does work, but if you plan to scale into the tens of thousands the return will be much slower as it will have to process it in batches of 500. Pinecone would allow you to literally scale to hundreds of thousands of submissions in an optimized way and offload the strain from your WordPress SQL database.

I was able to test and see the error. I believe what could be occurring is that it’s maxing out your PHPs memory limit with that amount of products. I was already reviewing and implementing a more optimized approach.

I’m going to optimize how the chatbot retrieves product data. Instead of loading all 2,700 products at once (which can overload the system), I’ll modify it to load them in smaller batches. This should prevent the system from crashing while maintaining the chatbot’s performance.

I’m still uploading data to the test site’s knowledge database – currently at around 1,800 entries. I’m going to load about 3,000-3,500 entries and will test it tomorrow to see if I can replicate the issue and confirm my fix works (which it should). Bear with me while I work on this, but I’ll get it fixed ASAP.

Yes, please add it back. I would like to review. I know it’s frustrating to add them all back and apologies for that. If you can ping me once it’s done I can check. If you’d like, you can disable the autoshow chatbot option and use the shortcode on a test page for now while we check for the issue. That way you can leave the bot accessible to me but not to your visitors.

You can use the shortcode: [mxchat_chatbot floating=”yes”]

If you do this, just send me the URL here and I’ll take a look once it’s back imported.

• This reply was modified 1 month, 3 weeks ago by mxchat.

Hi @tomchen29 ,

Thanks for reaching out. I want to help you with the issue. There shouldn’t be any issues with the amount. While I’ve personally only tested up to around 2,000, one user has told me they’re using it with around 22k pages submitted. I went to the URL and tried to test and it seemed to work. Were you able to resolve it?

I see your edit now – it appears you deleted the knowledge. If you are able to add it back I could check the frontend to see any potential errors. Meanwhile, I will push one of the websites I have up to 3k for testing.

• This reply was modified 1 month, 3 weeks ago by mxchat.
• This reply was modified 1 month, 3 weeks ago by mxchat.
• This reply was modified 1 month, 3 weeks ago by mxchat.