Видео

I believe the agent node is not able to trigger the tool nodes connected to it or there are issues with the tool workflows where the response will not go through the main agent node. Make sure to test each of the tool nodes before deploying to telegram, to ensure a response gets through to the AI Agent node.

@@limitlessai_media Hello, thank you for your quick reply. Unfortunatelly all tool workflows are connected to the Ai agent, google options to categorize documents are also working. However, if I write "Hello" the same happens. Tools are not working so I might have to redo the Ai agent. thank you and will send my solution.

Thanks for watching the video. Please check the links on the description to download all resources available.

@@limitlessai_media if I downloaded all of the resources, how to insert the n8n JSON template into the flow? Do you have a link to teach me how to insert the template?

@@mynas-ng5tt I haven't had a dedicated video on importing the template to your n8n workflow. But I gave a step-by-step guide from the Download page so you can start using the templates within your own n8n account.

Definitely capable, it will take a bit of code though. Defeats my purpose of providing low code to no code AI builds, reason I did not dive deep into this implementation. If you did this within n8n, it would be awesome if you can share with everyone 😊

Thanks for watching the video @vitalis! For this specific demo build there really is no advantage in using airtable over google sheets since we're just getting a minimal amount of data and google sheets would be enough. I made use of Airtable in this tutorial to show what's possible when building this project for a production environment where scalability and stability is of the essence for the applications you are going to build. Learned building these flows through messing around and being passionate about AI.

@ looking forward to see this channel grow. I’ve been looking for airtable alternatives with AI integration for automation but there’s only baserow. Upcoming are nocodb and teable. I’m now trying Rows, but agentic workflows are the next level.

Glad you find this video valuable! Great suggestion, yes it's possible but I'm afraid a complete AI SDR end to end build might not fit into 1 video. Fundamental build maybe? I'll look into this type of content soon! Thank you man.

a 1000X in seconds might not be possible, you are relying on the tokens/min response rate from an LLM. The data scraping part is good for faster and almost real time execution, but for the part where analytics and visualization are executed, LLMs are not able to keep up as they have rate limits imposed in a per minute basis. Doable if you scale into production level set-up, you need enterprise privilege to have higher rate limit of token utilization per minute.

For personal use maybe? The capability to scale and customize how the system fits to your needs? For data security? Or, some real-world application where data is supposed to be protected and not let some third party big company like OpenAI handle your organization's data? Just some of my 2 cents, but well, there's always ChatGPT and you can do whatever you like man. Anyways, thanks for watching the video, appreciate it so much.

I would love to if only n8n has the integration for E2B sandbox interpreter tool already. This amazing tool has just been available in Flowise. This video is meant for no-code/low-code people, If you do this in n8n you might need code solution to integrate a sandbox code interpreter tool like E2B and access it by an agent using n8n. Doable but requires you to code.

@limitlessai_media one should be able to implement it with api. Import the "curl". I am no coder, but is pretty easy to set up.

Thank you @SoloJetMan, glad you liked this video! Hope you got a few bits of value out of this tutorial.

You're awesome as well! Hope you get valuable bits on this tutorial. Happy building!

Can you add Groq somewhere in there to enhance the speed? Also, in custom tools where an api key or bearer token is required, where do I add env variables in flowise cloud? Or do I hardcode in JavaScript in custom tool node please?

@@MH-xx6df Groq or Llama for speed is a great suggestion, however, my problem there is the reliability of the coding prowess which is the main point of this video. As explained in the video, I used Claude latest since it is perfect for code use in data analytics and visualization. Haven't tried setting up an env for credential tokens yet, but I guess it is doable since Flowise and n8n is meant for production use. Great idea though, I'll look into this and maybe create some video for this types of solutions in future.

@@limitlessai_media cool and yeah, certainly get all the coding capabilities of 3.5 Sonnet. Use it all day long. Just thought might be away to have Groq help. Sorry, was thinking out loud.

Hey man, thanks for the feedback. I intentionally skipped the postgres chat memory since you can go along with the set-up using the basic window buffer memory option. I'll take note of this and will try to cover setting this up in my next videos.

@@limitlessai_media thanks I did it with that, still not able to run the app, but ive learnt a lto thanks

Thanks for your feedback. Personal finance is just a simple example of an accounting process. Basic accounting involves recording financial transactions, so this AI demo can already be considered an accounting process, imo. The video’s purpose is to show how AI automation tools can simplify such tasks. Building a fully automated AI accountant would require more time and advanced API integrations, which go beyond a low-code solution. However, it’s certainly possible with the right approach. If you're curious about AI's potential in this area, I'm happy to discuss it further.

Hi, thanks for your feedback. I see you're looking for the value in the Name field for the vector store node. You can assign any name you want for this field, but I suggest to name it the same as the INDEX name from your vector store of choice (Pinecone in my case). Regarding the error, I'm not familiar with the current limitations when building from a local install using docker, I'm using a self hosted version using render cloud myself and I haven't encountered any problem. Anyone here in the chat who might be familiar around this?

Yes of course, the same automation can be replicated within n8n.

Thank you for pointing that out, glad that you find this video useful. The dates and balance was a basic calculations in Google sheets which I think can still be improved. This video is to showcase the power of n8n, reason I did not focus on the sheets process too much. All that being said, appreciate your comment and will definitely take note of this in future videos.

@@limitlessai_media thanks

Glad that you find this content useful man, more to come!

The pleasure is mine! Glad this simple video made it on your timeline and hopefully you've gained a bit of value. Thanks for watching! Looking forward to connect with you 😊

Thanks for the question, glad you brought this up. Imo, creating GPTs with Retrieval Augmented Generation (RAG) just like the Learning companion we built in this video directly in Chat GPT will likely be more costly than using platforms like n8n since you're likely going to directly use OpenAI's cloud service to host and store your documents for retrieval. Also, flexibility in designing the behavior of your RAG agents is not fully embedded with open AI, that's why automation platforms like n8n is a best fit for this specific purpose. There are more I can discuss about the pros and cons of directly using ChatGPT vs platforms like n8n. If you are interested to learn more, let us connect through the links I provided in the description 😊.

Bro! thank you so much. I'm screenshotting this 😊

Thank you for the video, and well done! I like the format and appreciate the clear instructions.

@@MartinusBiemond glad that you liked it, appreciate this feedback so much as it gives me idea how to improve my creation process to give you the best possible way to learn more about AI use-cases, will definitely do more of this style and format on my videos soon!

Hey man, thank you and appreciate this very much! The entire templates are available in the description for you to download for FREE! Try it out yoursefl, enjoy 🍻🍻

Hey man, what specific AI workflow do you need help with? If you need any assistance for any project you have, please reach out to me on the links provided in the description.

Hey there, you can download the tools from the link provided (under Bonus content). Although not all of the tools shown in the video are included, but I hope it can already help you try out the capability of this AI Assistant for yourself after setting up on your own.

@@limitlessai_media doest the others tools are gonna be available soon ? ;) if no how can we work together ?

@@pienik.delrieu you can reach out to me on my contact details page via this link👉: limitlessai.digital I'm happy to share with you the resources you need. Also, looking forward to work with you!

The latest Claude computer use capability is totally different from this since the Claude demo is using code in the background to run the LLM enabling it to interface with your computer in a sandboxed environment. More like an AI agent running in the background that's powered by the Claude AI model guiding the agent to mimic mouse clicks, use tools, and execute computer commands. This is all I think I know about this new breakthrough from Anthropic, anyone here who's got more in-depth idea how it works? Feel free to comment!

On-prem, slightly yes since you can self-host using 3rd party cloud service provider like render and run the open-source instance of n8n giving you with more flexibility and control over your own data. However, if you want this to work offline, it can be possible by running locally at your computer, but you may lose the capability of scaling for organizational level usage where more people can interact with the AI Agent. Also, if you want greater privacy over your data, this might be trivial since the best LLM models that powers this AI agent to work as demonstrated are considered 3rd party entities. Let me know what you think though, or if someone has better take on this privacy-related topic. If you're interested to learn more for your own projects or business application, hit me up and let's discuss, I'll be more than happy to help🙂👉: calendar.app.google/ZY8cqGnHQbes7Mjc7

@@limitlessai_media happy to hear this. We are dealing more with Gov sectors that requires more privacy environment. Will explore more on this. Thanks for the reply

@@dayakrider349 If you need any help putting up the kind of solution you have in your mind, I'm happy to help! Reach out to us HERE👉: limitlessai.digital

Hey, @elfmun, please connect with me through the booking link I provided in the description. Otherwise, please look for my contact information in my Official Page👉: limitlessai.digital or book a FREE consult or discovery call HERE👉 : calendar.app.google/ZY8cqGnHQbes7Mjc7 And thank you for watching my video!

Can we get step by step build maybe not so may tools but showing you actually building the bots it doesn't have to flashy just you doing it

@@sandowssecret7363 Thanks for that awesome suggestion, watch out for that video soon!

Filipino brother, glad you made it in this video! Hope you enjoyed and got some valuable bits.

Honestly I haven't tried it yet but I think I found some video showing that it's possible to implement multiple language. Not sure, I'll be checking on it, thanks man!

What works for me to address hallucination is tons of trial and error in testing the combination of prompts and the AI model to use until I get the result I wanted. OpenAI 4o model works best for me when building AI Agent workflows or any tool calling implementations, while I use Claude for RAG projects. Anything that interacts with documents like retrieval, and when you want the responses to be more grounded with your specific data, Claude models seems to be more accurate for this types of use-case. In the end it all depends on how you structure your prompts, and test rigorously with your AI model of choice 🙂

medical records being processed through OpenAI (or any public model)...not a good idea.

@@0xb1sh0p8 Totally agree, not unless you use the enterprise versions of those public model since when it comes to big businesses, these model providers tend to implement safeguards and compliance guardrails around their products if dealing with enterprise accounts. Not sure though haven't looked at it yet. What are your thoughts though?

Yes, this AI assistant is deployed using Telegram chat. You can definitely interact with it from your mobile phone.

@federicoarg00, the pleasure is mine! Glad that you liked this. Let me know if you have anything in mind around this topic that you want me to cover in my next videos.

@ramielkady938, I'd love to know if you're interested to learn more about n8n. Let me know what specific application or part of the process you'd like me to tackle next 🙂

@@limitlessai_media The part tells you how, how much does it cost to subscribe.

@@Rami_Elkady the nice thing about n8n is that you have an option to self host it and run locally from your computer or use third party cloud hosting services. Or you can pay $20 monthly for cloud account. I personally use Render basic tier plan which cost me about $17/month. With this, you can already build & deploy unlimited number of workflows and automations with all the capabilities and tools n8n has to offer. The n8n source code can be deployed at no cost using the open source repo available at GitHub. I'll probably make a video soon on how you can do this step-by-step.

Possibilities are limitless, more to come!

Thanks man, glad you got a glimpse of this powerful AI tool. Possibilities are limitless when building with n8n.

I'm glad that you enjoyed the video, let me know if you have something in mind you want me to implement AI with in future videos. I look forward to a successful collaboration.