What is Flowise AI?
Hey, what is up guys? Welcome back to another YouTube video at the world of AI. In today's video, we're going to be focusing on an amazing new application which is called Flowise AI. Now, Flowise AI is a platform that offers a drag-and-drop user interface for creating customized workflows using LangChain.js. Now LangChain.js is a programming language and a framework designed for building natural language processing applications.
Now by combining the ease of use of a drag-and-drop interface with the power of LangChain.js, Flowise AI enables users to create sophisticated language processing workflows without the need of extensive coding knowledge. Now with the revamp Flowise API, users as well as developers can build a front-end application that provides users with the ability to upload various types of documents and websites. These uploaded resources can be processed and analyzed using customizable workflows created within Flowise AI.
Now the APIs are allowing developers to interact with the Flowise app platform by enabling seamless integration of language processing capabilities into their own application. Now in this demo, you can see you're able to combine different applications as well as different LLMs to create a functional as well as a conversational chatbot. Now this is an easy way for you to utilize different tools of LangChain as you're able to plug and play into different apps as well as tools to get you a conversational chatbot as well as a different type of app that you can build off of LangChain. And this is something that we're going to be focusing on today's video.
Community and Engagement
Before we actually get into the gist of the overall video, if you guys haven't followed my Twitter page, please do so guys, as there's going to be a lot of content as well as a lot of news that I'm going to be dropping on this Twitter page, so definitely give it a follow. Definitely subscribe, like this video, and comment anything you want to see in upcoming videos. I'm going to be posting a lot of content and a lot of video that will definitely benefit you guys, so definitely do so. And with that thought, let's get right into the video.
Key Features: Filtering and Chaining
Now one of the key features of Flowise AI is its ability to filter uploaded documents and websites based off of metadata and namespaces. Now metadata refers to the additional information that is associated with the uploaded resources, such as like an author's name or publication date or even relevant attributes that correlate to that uploaded document. Now namespaces, on the other hand, provide a way to organize as well as categorize resources based off the different criterias. Now by leveraging the filtering capabilities, you as a user can actually extract specific information from the uploaded documents and websites based on the provided metadata or the namespaces. This allows for targeted analysis and extraction for relevant data which can be used for a variety of different applications as well as content creation, information retrieval, or for data mining.
Now this is one of the main features that they've actually emphasized with the Flowise application, but there's also different things that you can do. You're able to chain different LLMs, you're able to build off of them and create a Q&A retrieval chain, you're able to even have a language translation chain that is correlated within the application. And this is something I'm going to be showing you as to how you can download, and we're going to go a little bit more in-depth as to what you can do with it.
If you actually go on Flowise AI, you can see that there's different things that you can do. You're able to visually chain your own LLM with an AGI on your own data. Here you can see these different examples on their Twitter page and you can see what people have been able to do as well as build on top of the Flowise drag and drop user interface. And now with that thought, let's get on to the next step of the video where we actually show you on how you can install it onto your desktop locally.
How to Install Flowise on Your Local Machine
First things first guys, you're going to need Git, which is going to be used to clone the actual repository onto your desktop. Secondly, you will also need Python, the latest version to install it as well as use it as your code editor on your desktop. So with that thought, first things first, what you want to do is go up, go click the screen button on the repository and copy this link. I'll leave all the links in the description below so you can get a better access for them.
Now what you want to do is open up command prompt. Once you have done that, you want to type in git clone and paste the link, and what you want to do is click enter. And once that is done, you want to click cd flowwise and this will basically put you into the Flowwise folder. Once you have done that, you want to install the packages onto your desktop. That is by copying this command over here and pasting it on your command prompt and press enter. And now this might take a couple of minutes. Once it's done, I'll be right back to you and then we'll be able to host it on our localhost by starting it on with this prompt over here. And with that thought, I'll be right back once it's finished installing.
Once it has finished installing all the packages what you can do now is copy this code and you can paste it right over here and click enter and what it'll do is it will start hosting on your localhost and you just got to click on this link over here and it'll take you to the localhost. Now once it's finished uploading it will be ready to upload over here. You just gotta click refresh after a couple minutes and it will be uploaded on your local host.
Exploring the Flowise User Interface
And just like that it is uploaded onto our Local Host. Now before we actually get into the actual just as well as the features of what Flowise can do, I want you to first turn on the actual dark mode. You can play around with it with your own preference over here. You can load data spaces as well as export databases, which could be useful in a lot of different use cases.
Now the chat flow is where you're going to be working around the most as this is the place where you'll be playing around as well as chaining different LLMs or whatever applications you want to integrate and build. Over here you have your panel you can play around with. In the API Keys, you can upload your different API Keys you want to create or upload. Obviously, you can also upload it on the actual chat flow, which I'll show you later on.
Now the marketplace has different types of actual applications that are already built. You can see over here you have different things like a translator, web browsing chatbot you can create using LangChain. You have a multiple vector database, you have a GitHub repo Q&A bot. You actually have BabyAGI as well as Auto-GPT which is quite useful for a lot of different people who want to use autonomous agents and chain it with different LLMs to help you get the best results. So this is also another cool thing that I'll be showing you guys as to what you can do. There's also a ChatGPT plug-in feature, metadata filter upsert which is something we talked about before, but in this case we're just going to go with something easy which is a conversational agent. So in this case, we'll just click it and we can use this template right here. Once you have done that you can input your SerpAPI, your open API key, a calculator which is used to actually help your buffer memory. In this case, you can use Pinecone, so you can even swap it for a different one and your conversational bot will be able to have conversations on this chatbot over here. So this is how you actually upload and execute a certain workflow. You can even rename it as well as tweak around with it later on.
Building Blocks and Potential Applications
In this case I'm not going to be like showing you this one I was just showing you just how the workflows would actually work. In terms of the actual chat flow, you can come on over here, add a new one, and you can click on this button over here and you can add different nodes and different applications. There's different agents you can even play around with which are over here. You can have different prompts, have different prompt templates, set up a vector storage, you can integrate like Pinecone for example to store your search term or long-term memory. You can have different tools that are utilized in the actual LangChain toolsets. You have different things, LLMs as well that you can utilize.
And this is just a start as to what they're going to be adding guys, as they're going to be continuously working towards adding new features as they've been actually working in on adding different things as well as different features that will enhance this overall application. And the great thing is guys, this is a free tool that is open source and it's used for commercial as well as personal use at zero cost. And the best thing is it's built off of LangChain.js and in this ecosystem of LLMs, you're able to have various building blocks which can be used for different use cases and this is something I'm going to be showcasing you. Whether if it's an engineering prompt engineering feature, you can utilize agents, chaining, semantic search, chat models, vector storages, and other tools that could be assigned to an agent for specific actions as well as functionalities. Now this is something I'm going to be showcasing later on in the video.
But now let's actually get on and focus on some of the features that you can utilize this overall Flowise application. Now we already know that you can do things such as chaining LLMs, but with the text classification feature, you can build an LLM chain to classify text documents into predefined categories or labels. Now this is useful for applications like spam filtering or sentimental analysis. Another feature that I wanted to emphasize is its named entity recognition. You're able to develop an LLM chain which you can extract named entities using like personal names or organizational names or even dates that could be applied to extract as well as use them to basically export different types of databases within recognition of certain types of names. Now this is just one of the two features that I wanted to talk about, but let's actually show you some of the things that you can do in terms of building different apps using this application.
Demo: Creating a PDF Q&A Bot
Alright guys, so I took the time to actually create a no-code PDF Q&A bot. Now this is something that's absolutely amazing. It legit just took me one minute to sort this all out. So what I did is I had a recursive character text splitter and from this it inputs it into the PDF file. Now from the PDF file, I used an OpenAI embedding which connects the actual Pinecone storage, which retrieves as well as inputs all the storages into the actual Pinecone. And what it actually does with the ChatOpenAI is that it uses it as well as utilizes the tools to actually chat with the PDF file. So I just actually created a Q&A bot for chatting with your own PDFs and files, and it's easy by inputting these different things.
For example, if I want to keep a conversational retrieval Q&A bot, I just put it in like that. Obviously I already inputted all the things and where it says to input the API keys, I just went on to my API keys and inputted it over there as well as my Pinecone storage and input it over there. So what you need to do is you need to connect each of them by chaining them. So what you can do is get the spots or the dots over here and connect it to the different documents. Now the document goes to the document, the embedding goes to the embedding. So in this case embedding goes over here and in terms of OpenAI, ChatOpenAI, the language model gets connected over here, and the vector storage gets connected right over here. And just like that you can then start chatting with it. But in this case I could actually upload a file, so I'm going to upload this first and I'll be right back. It might take a couple seconds, but I'll be right back.
All right it seems like it's uploaded and once it's uploaded you just have to click save and you can just say PDF chat bot Q&A all right and then once that is done you can okay no not don't save that. Now the workflow is saved and you can now start chatting. What does the PDF entail? Okay, I don't know what this error is but I'll be right back.
Troubleshooting and Final Thoughts
All right guys, so I can't seem to figure out the error or what's happening. I think it's because of my Pinecone account, the key is not inputted properly. I don't know what is happening, but basically you're able to get and start chatting with the bot. There's actually a video on how you can, it'll teach you how as to how you can do that, so I'll leave that in the link in the description below so that you can get a better idea as to how you can actually play around with this PDF chatbot.
But this is just one example as what you can do. You can also do the BabyAGI bot. I've also been playing around with the Q&A bot, trying to talk to it. So I actually asked it, what is LangChain? I'm able to get this response, and you can see that you can do a lot of different things and create a lot of different bots as well as different chains that can help you out with a lot of different things. So I definitely see this as a very innovative tool that provides you a lot of different things as it's an easy way for you to deploy different tools and agents from LangChain.js. And you can use this tool to do a lot of different things. So this is quite amazing as to what you can do with guys.
So in summary, I definitely see that Flowise provides such an easy interface for developing LLM applications and I see this as a very cool tool that can be used for a lot of different use cases. So I hope you found this video quite useful and it was a lot, uh, you got some sort of value out of it guys. With that thought, thank you so much for watching. Make sure you do subscribe, turn on the notification bell, like this video, and comment anything you want to see. And make sure you definitely follow the Twitter page as you'll get the best content. If you guys haven't seen any of my previous videos definitely do so as there's a lot of knowledge and value that you'll definitely benefit from. And with that thought guys, I hope you found this video useful and with that thought have an amazing day, have a bright smile, and I'll catch you very shortly. Peace out fellas.
