Thanks a lot, Leon, you are basically the leader in UA-cam when it comes to Flowise learning. I really appreciate every minute you put in to help us in this journey. Will you create a video on how we can build AI persona chatbots? Who have functions like - Effective QnA - Link retrieval - Strategy recommendations?
@@leonvanzyl Absolutely. I watched the whatsapp video and thought I would try telegram as I only have my personal number on whatsapp and didn't want a bot on that
I appreciate your free tutorials (keep it up!) but, I'm sorry, for this one I have to give you critical feedback for misleading people. This workflow is missing the key component-saving and passing on the GPT Assistant thread ID. Without the thread ID-the assistant responds to every message as if it's the first message in the conversation without being able to take the conversation context into account. Such assistants are pretty useless without such functionality. Nevertheless, I'm grateful to you for showing how to overall connect to Telegram.
@@leonvanzyl , yes! What I found out afterwards is that Flowise allows to pass on any value as the Flowise chat ID by just updating the HTTP POST Request content to: { "question": "{{1.message.text}}", "chatId": "{{1.message.chat.id}}" } This makes it so much easier than working with GPT Assistants directly as GPT Assistant thread IDs can be created just by OpenAI itself based on their pre-defined ID structure (meaning that you need to establish an extra database matching GPT Assistant Thread IDs with chat IDs from your channels (e.g. Outlook Email Thread ID, Messenger Sender PSID, Telegram Chat ID etc.). So this Flowise feature alone is gold. Thanks once again for your tutorial!
@@leonvanzyl Leon I too have run into the same problem. When asking the bot questions in telegram, it treats each message as a new dialog. And then it turns out that the Telegram bot is absolutely useless. Please tell me how to fix it. Please give me recommendations, or better yet, a step-by-step instruction.
Hi Leon, first of all I'd like to say thank you. As I'm a not a tecnical person, this videos are giving very powerful tools to make useful applications. I'd like to ask you a question about this video. I'm following your video step-by step and when I try to add the zoho node in my make application it doesnt show the same fields that appears in your video, actually it doesnt show any fields at all! Just the dropdown connection and the "show advanced settings" do you know whats happening? Do I have to make any configuration in Zoho before applying in make platform? Once again, thanks a lot!
I appreciate your free tutorials (keep it up!) but, I'm sorry, for this one I have to give you critical feedback for misleading people. This workflow is missing the key component-saving and passing on the GPT Assistant thread ID. Without the thread ID-the assistant responds to every message as if it's the first message in the conversation without being able to take the conversation context into account. Such assistants are pretty useless without such functionality. Nevertheless, I'm grateful to you for showing how to overall connect to Telegram. Leon I too have run into the same problem. When asking the bot questions in telegram, it treats each message as a new dialog. And then it turns out that the Telegram bot is absolutely useless. Please tell me how to fix it. Please give me recommendations, or better yet, a step-by-step instruction.
You're 💯 correct. Sometimes I need to be selective in my tutorials on what I can realistically include without the video becoming to long. If there's a demand, I'll gladly do longer form videos. I would definitely store and retrieve threads based on the user 👍
Leon and here in the comments can you tell me how to fix this problem. Not later in the next videos. And in the comment in text tell me how to fix this error. I think it will be useful to many people who will do further on this lesson. I will be very glad if you answer here how to fix this error. @@leonvanzyl
Thank you 🙏. You can build some pretty cool chatbots with Flowise. It can add some custom functionality to OpenaAI (like the lead generation example I showed at the beginning). Also, many people would want to use Open Source models for this, especially for roleplay and persona chatbots. Again, this is very easy to build with Flowise.
How can we add a stripe paywall as well as adding saving and passing thread Id so it can run indefinitely to potential customers and cushioning the costs as well?
I doubt that you'll be able to add a paywall within the chat prompt itself. It is definitely possible reuse the same thread, but the setup for this is quite complex. We could still use Make.com for this, or something like Voiceflow, where we first fetch the thread for the user from a database (like Airtable), and if a thread does not exist, we first need to create it. And then pass the thread to Flowise / OpenAI.
I really like your content and the way you explain everything... I am stuck in deploying flowise on AWS EC2... Can you please make a video on that or can you please guide me... I really appreciate it... Thanks
Leon thank you for your videos, they are very useful. And a big request, please make a video instruction how the Flowise bot transmitted customer data to Telegram and Email. I will be very grateful to you if you make such a video.
It is possible, but incredibly complicated as it involves calling multiple API endpoints. Flowise makes it easy since we only have to call a single endpoint.
Great thanks a lot. Do you plan to make more Videos about Flowise? Can you maybe show how to Update Flowise in Amazon AWS or in general? Would be great!
You can definitely use different models. Although function calling is still limited to OpenAI, but you can still answer questions from vector stores using Open Source models.
This guy is awesome. what a fluent way he does everything. Always wait for the next tutorial. Thanks Leon
Thank you very much for the compliment 🙏
Thanks a lot, Leon, you are basically the leader in UA-cam when it comes to Flowise learning. I really appreciate every minute you put in to help us in this journey.
Will you create a video on how we can build AI persona chatbots? Who have functions like
- Effective QnA
- Link retrieval
- Strategy recommendations?
Thank you very much for the kind words 🙏
Hi Leon, I know n old video but awesome, I just did this using n8n insead of make, works like charm. Thank you
Hey Wayne! That's awesome! n8n and Flowise is definitely a powerful combo
@@leonvanzyl Absolutely. I watched the whatsapp video and thought I would try telegram as I only have my personal number on whatsapp and didn't want a bot on that
Silly question, will the whatsapp bot override your normal number or will it be different?
I appreciate your free tutorials (keep it up!) but, I'm sorry, for this one I have to give you critical feedback for misleading people. This workflow is missing the key component-saving and passing on the GPT Assistant thread ID. Without the thread ID-the assistant responds to every message as if it's the first message in the conversation without being able to take the conversation context into account. Such assistants are pretty useless without such functionality. Nevertheless, I'm grateful to you for showing how to overall connect to Telegram.
Ah yes, you are right. This solution can definitely be enhanced to store and pass around the Thread ID as well 👍.
@@leonvanzyl , yes! What I found out afterwards is that Flowise allows to pass on any value as the Flowise chat ID by just updating the HTTP POST Request content to:
{
"question": "{{1.message.text}}",
"chatId": "{{1.message.chat.id}}"
}
This makes it so much easier than working with GPT Assistants directly as GPT Assistant thread IDs can be created just by OpenAI itself based on their pre-defined ID structure (meaning that you need to establish an extra database matching GPT Assistant Thread IDs with chat IDs from your channels (e.g. Outlook Email Thread ID, Messenger Sender PSID, Telegram Chat ID etc.).
So this Flowise feature alone is gold.
Thanks once again for your tutorial!
@@leonvanzyl Leon I too have run into the same problem. When asking the bot questions in telegram, it treats each message as a new dialog. And then it turns out that the Telegram bot is absolutely useless. Please tell me how to fix it. Please give me recommendations, or better yet, a step-by-step instruction.
Hi Leon, first of all I'd like to say thank you. As I'm a not a tecnical person, this videos are giving very powerful tools to make useful applications. I'd like to ask you a question about this video. I'm following your video step-by step and when I try to add the zoho node in my make application it doesnt show the same fields that appears in your video, actually it doesnt show any fields at all! Just the dropdown connection and the "show advanced settings" do you know whats happening? Do I have to make any configuration in Zoho before applying in make platform? Once again, thanks a lot!
You actually need to trigger the flow once in order for the data to appear.
I appreciate your free tutorials (keep it up!) but, I'm sorry, for this one I have to give you critical feedback for misleading people. This workflow is missing the key component-saving and passing on the GPT Assistant thread ID. Without the thread ID-the assistant responds to every message as if it's the first message in the conversation without being able to take the conversation context into account. Such assistants are pretty useless without such functionality. Nevertheless, I'm grateful to you for showing how to overall connect to Telegram.
Leon I too have run into the same problem. When asking the bot questions in telegram, it treats each message as a new dialog. And then it turns out that the Telegram bot is absolutely useless. Please tell me how to fix it. Please give me recommendations, or better yet, a step-by-step instruction.
You're 💯 correct.
Sometimes I need to be selective in my tutorials on what I can realistically include without the video becoming to long.
If there's a demand, I'll gladly do longer form videos. I would definitely store and retrieve threads based on the user 👍
Leon and here in the comments can you tell me how to fix this problem. Not later in the next videos. And in the comment in text tell me how to fix this error. I think it will be useful to many people who will do further on this lesson. I will be very glad if you answer here how to fix this error. @@leonvanzyl
very useful video...
why u deleted openai module and added flowise module, how does it effect tokens cost back and forth from models used.
Thank you 🙏.
You can build some pretty cool chatbots with Flowise. It can add some custom functionality to OpenaAI (like the lead generation example I showed at the beginning).
Also, many people would want to use Open Source models for this, especially for roleplay and persona chatbots. Again, this is very easy to build with Flowise.
@@leonvanzyl oh ok i thought the reason of deletion was that Flowise is opensource and free to deploy while openai module is not.
Absolutely love the videos as usual! Is there any way to add multiple choice selections/buttons in a chat flow?
Thank you 🙏.
Unfortunately not at the moment.
How can we add a stripe paywall as well as adding saving and passing thread Id so it can run indefinitely to potential customers and cushioning the costs as well?
I doubt that you'll be able to add a paywall within the chat prompt itself.
It is definitely possible reuse the same thread, but the setup for this is quite complex.
We could still use Make.com for this, or something like Voiceflow, where we first fetch the thread for the user from a database (like Airtable), and if a thread does not exist, we first need to create it. And then pass the thread to Flowise / OpenAI.
I really like your content and the way you explain everything... I am stuck in deploying flowise on AWS EC2... Can you please make a video on that or can you please guide me... I really appreciate it... Thanks
Thank you!
For interest sake, what issue are you having with EC2?
@@leonvanzyl
I am getting the following error:
TypeError: kwargs_from_env() got an unexpected keyword argument 'ssl_version'
Leon thank you for your videos, they are very useful. And a big request, please make a video instruction how the Flowise bot transmitted customer data to Telegram and Email. I will be very grateful to you if you make such a video.
Great idea, thank you 👍
Many thanks for the great work Leon! Is it possible to integrate an assistant API build in the playground using this method? 🙏🙏🙏
It is possible, but incredibly complicated as it involves calling multiple API endpoints.
Flowise makes it easy since we only have to call a single endpoint.
Great thanks a lot. Do you plan to make more Videos about Flowise? Can you maybe show how to Update Flowise in Amazon AWS or in general? Would be great!
Absolutely, I'm actually creating an updated series on Flowise. A lot has changed since I released the first crash course.
Why do I get this error "RuntimeError [400] Bad Request: chat not found" is it possible to help?
That means the Flowise chatflow does not exist.
sir can we do it with gemini pro or any other opensource llm?
You can definitely use different models. Although function calling is still limited to OpenAI, but you can still answer questions from vector stores using Open Source models.
@@leonvanzyl thanks
Can i connect Instagram?
Nice job, mate! Are you considering to teach WhatsApp deployment too?
Thank you!
For sure, just trying to find an easier way to do it.
Amazing@@leonvanzyl Really looking forward to WhatsApp integration, thanks 👍
amazing! thanks man
Thanks
Thank you very much 🙏
Love it
You're welcome
awesome!!
Thank you 🙏
I mean stop using no code pls, it's suck