What is FlowiseAI, and why should I use it?
First, let me start by saying that I am not an expert programmer; I like it, but I do not write code for a living. For me, it’s a way to earn extra money (maybe one day), be creative, learn new technology, and network. With that said, let’s dive in.
So, what is FlowiseAI? From the Github page, FlowiseAI is a “Open source UI visual tool to build LLM apps”. That means it’s a no-code/low-code way of building LLM apps using a visual platform. It is straightforward to build because of how the interface works, and it’s more idiot-proof than other methods.
How do I use it?
It’s straightforward, and I will give you the easier answer here. First off, you will need Docker desktop (MacOS/Windows/Linux), this provides the required back-end for your OS to handle containers. Docker Desktop is a handy tool you should use (it’s free!), and learn how to use it. The biggest reason has to do with Python applications, not Python in and of itself, but because of how Python apps handle environments (Python versions, dependencies, libraries, etc.). One way to avoid any of this environment nonsense is to use the Docker version of the install. This is what I did; because of that, I had 0 issues with versions, libraries, dependencies, conflicts, etc. It was the “ez-mode” version of running FlowiseAI.
Once you have it up and running (follow the instructions on the GitHub repository; they are very straightforward) and are connected to the instance via localhost:3000, then it’s time to create your first chat flow.
What should I build?
Start with a simple chatbot for a website, as an example. I found this YouTube video the other day. The idea was a simple chatbot using Pinecone for the vector store, a web scraper to gather the data, and OpenAI to do the heavy lifting. It sounds simple enough, and frankly, it was. Follow along and build the same thing they did, and you will get the hang of it quickly. Or try to use a local company with a web presence and build the same thing for that retail company. If you have it working, you can approach them and offer to build it for them.
