How to Build a Q&A Chatbot with RAG (No Code!) - Langflow Tutorial #3

[@VikasSawantMumbai]

Leon, I want to thank you from the bottom of my heart. I have been playing on and off with GenAI with LangChain/OpenAI etc, but these 3 tutorials (#1, #2 and #3) were excellent and have clarified my rusty thought process and I am simply thankful to you the way you explained - not fast, not slow, very clear and interesting. I have subscribed and likes your video and will go thru the rest, earlier I thought I will go thru as many and then post my thanks note, but after tutorial #3 my happiness did not allow me to hold back. Thank you, very nice!!

[@rajnd]

Sincere Thanks Dear Leon for making these tutorials available for all. It makes understanding very simple. Appreciate it. Lotsa regards

[@leonvanzyl]

You're welcome 🤗

[@TomPooleMS]

Leon - your videos are so informative and presented so they are easy to understand. I cannot thank you enough.

[@leonvanzyl]

You're welcome 🤗

[@nufh]

So this is how RAG works! I always thought we just loaded the entire text into the context window. Thanks for explaining it so clearly; it really helped me understand!

[@leonvanzyl]

You're welcome 🤗

[@HyperUpscale]

Finally! A great explanation how the vector store works in LangFlow 😅

[@HyperUpscale]

Mystery solved - after one year ...

[@leonvanzyl]

Thank you 😊

[@HyperUpscale]

@@leonvanzyl Actually, Thank you!

[@WayneBruton]

Great series thus far Leon, I like the look and the fact its Python. I hope you go all the way with this series.

[@leonvanzyl]

Thanks man! Anything you would like me to cover?

[@WayneBruton]

@@leonvanzyl Agent Teams definately also, would love to know the strengths and weaknesses or a comparison between flowise and langflow. They both seem awesome. I guess ultimately it will be "horses for course" depending on your needs

[@BismaMajeed-si1uf]

your channel is a goldmine! thankyou @leon

[@leonvanzyl]

You're welcome 🤗

[@nicolaslemosribeiro8096]

Thank u man, u save my life! continue this serie i need very much

[@leonvanzyl]

Will do

[@myandrobox3427]

Another great video, thanks, Leon! Does Langflow support images as input and output generated images?

[@AnujSaxenaMunich]

A simple text based rag application is explained at various places but there is no explanation if we can create a multimodal RAG application with either Langflow or with Flowise. Is it possible to get an example showing the usage of images in both input as well as output?

[@jgz2]

Thanks Leon, again very helpful.

[@4pass115]

Sorry if the question seems simple, I'm not a programmer. After finishing the assistant/chat, how do I make it stay online. Do I need to store it on a server? Do you have a video showing this step? Even in flowise, you install it on the website, but if you don't access it from the computer where the "node" and flowise were installed, it gives an error.

[@AKASHROY100]

Is there any way to extract images from the knowledge base documents and retrieve them later in the output ?

[@julienechavidre6389]

Thanks a lot Leon for your video series. It helps a lot. Is there a way to have a video to produce simple code files (like JSON), based on "code samples" (probably stored in a vector store)?

[@bekakakachia147]

Hello Leon, how can we control sessionid, when exporting flowise workflow via API, in Telegram bot, so it can have unique session id with each user id? Thanks in advance ❤

[@alessandro.depoli]

Hi, I'm deciding which tool to learn to start my journey, but in the last few days, I've been switching between Flowise and Langflow with double effort. Checking also Vextorshift. I like a lot Langflow for the style and the fact that it is "already online" with the new v1. Which platform do you suggest to study and use? Thank you (thanks for all the videos!) Are you continuing the series on Langflow? And on Flowise & VectorShift?

[@doctordude5728]

I don't have a section called "core components" so I am missing "Inputs, Outputs, Data, Models.." actually everything except "Vector Stores" and "Embeddings". What I've got are only the components that you have under the section "Extended"! So how can I get all the other components that I need to create a RAG Chatbot? Can you explain that please?

[@leonvanzyl]

You need to install the pre release version of Langflow. Check out my Getting Started video for installation instructions 👍

[@leonvanzyl]

kzbin.info/www/bejne/jJPUl42gicSDgqc

[@doctordude5728]

@@leonvanzyl thanx a lot!

[@jonathanccalver]

Hi Leon. Thanks for all your videos. I've watched a lot them. Sometimes you ask for project ideas that might be useful. I've got one for you. I'm looking to design a RAG that can chat with CSV data. I've gone through the normal process, as per pdf, docx files, but the results were not good. I suspect a different approach is required, right?

[@leonvanzyl]

Thank you for the suggestion! Alright, I've added a CSV RAG chatbot to my backlog 😊👍

[@FacadeEngineering]

Thanks. Can you also make video on html output of langflow? It is not working great like flowwise. Also how to make embedding on database trouth api and customizing components are great topics

[@synergy-xl6mt]

Compare Flowise what's advantage of Langflow

[@maniecronje]

Apologies if I already ask this personal preference FW or LF? Do you know if LF have more features and control than FW? I have not use LF but would like To give it a go if it is better than FW in terms of features etc. would appreciate you own personal opinion thx 🙏

[@leonvanzyl]

I think Langflow has potential, but they're not as advanced as Flowise.

[@bekonrexha]

My vector search gives this error for some reason "ValueError: Error building Component Pinecone Search: Invalid inputs provided." any idea why?

[@TarabUTK]

Where can I find the menu document download? Awesome video. Thanks

[@leonvanzyl]

Thank Mark! I'm not familiar with such a node. Are you trying to download the flow itself? Or something else?

[@jankotyza8146]

Hi, The video is well-made, but I'm having trouble getting it to work with Chroma or Pinecone. It seems that Langflow has changed since the video was made. For starters, there is no "Vector search" menu as mentioned, only "Vector Stores." Additionally, both Chroma (now Chroma DB) and Pinecone have different connection points on the right side compared to the video. There are now Retriever, Search Results, and Vector Store connection points, none of which can be connected to the "context" variable on the Prompt node. As a result, I'm stuck halfway through. Perhaps the video is outdated and could use a revision? Thanks!

[@leonvanzyl]

Yeah, super frustrating for us creators when the platform changes just about everything with each release 😂.

[@abedmalak_]

Why langflow? Flowisee is better, right?

[@professions_]

A question about Chroma please. Can I add different types of data files (pdf, doc, json) using the Chroma local storage option, or will it overwrite the data?

[@leonvanzyl]

Yes, you can add multiple files (and different file types). The content is ADDED to the vector store.

[@TarabUTK]

Hello, is video #4 available for pro members?. I am not able to see. it says"1 unavailable video is hidden" Thanks.

[@leonvanzyl]

Hey there, not yet. Still working on video 4 and 5 which will both release early for you next week 😊

[@josielgunner]

These search components don't appear to me, were they removed?

[@leonvanzyl]

Are you on the latest version?

[@josielgunner]

@@leonvanzyl Yes, although I noticed You used a Parse Data component instead on Rag with Ollama video. So, I used this same component on this chat and worḱs.

[@professions_]

Hi Leon, aside from Deploying on Google Cloud Platform, is there another way to host Langflow online? I love your tutorials! Thank you

[@leonvanzyl]

I'll see if I can cover something other than GCP as well 👍

[@cyberir]

Maybe free oracle cloud