Create an AI Agent that Writes & Runs Code in FlowiseAI (No Coding Required)

Flowise AI (2024) TutorialAging📅 2024-10-28

Build Process

┌─────────────────────────────────────────────────────────────────────┐
│                    BUILD CHATBOT WORKFLOW                            │
├─────────────────────────────────────────────────────────────────────┤
│                                                                     │
│   PLANNING                                                          │
│   ────────                                                          │
│   ┌─────────────┐    ┌─────────────┐    ┌─────────────┐            │
│   │   Define    │───►│   Choose    │───►│   Setup     │            │
│   │   Scope     │    │   Stack     │    │   Project   │            │
│   └─────────────┘    └─────────────┘    └─────────────┘            │
│                                               │                     │
│                                               ▼                     │
│   DEVELOPMENT                                                       │
│   ───────────                                                       │
│   ┌─────────────┐    ┌─────────────┐    ┌─────────────┐            │
│   │   Build     │───►│   Add AI    │───►│   Test &    │            │
│   │   Features  │    │   Logic     │    │   Debug     │            │
│   └─────────────┘    └─────────────┘    └─────────────┘            │
│                                               │                     │
│                                               ▼                     │
│   DEPLOYMENT                                                        │
│   ──────────                                                        │
│   ┌─────────────┐    ┌─────────────┐                               │
│   │   Deploy    │───►│   Chatbot   │ ◄── LIVE!                 │
│   │   to Cloud  │    │   Running   │                               │
│   └─────────────┘    └─────────────┘                               │
│                                                                     │
└─────────────────────────────────────────────────────────────────────┘

Key Components

Agent configuration
Tool integration
Memory management

Transcript

[00:00] In this video, I will show you how to use the E2B Code Interpreter in Flowwise. A Code Interpreter allows our AI agents to write and run code to do some pretty cool things, like analyzing complex data and visualizing the results. For instance, we could ask our AI assistant to analyze some data and generate some graphs. So from the chat window, we can select our file containing financial information about several companies, and the agent will be able to respond with any

[00:30] type of chart that we want. Like the stock price versus gross margin scatter plot, working with complex data and visualizing the results is a very common use case for using code interpreters. But you can definitely do way more advanced things with this. So in this video, we'll use Flowwise AI, which is a free no-code AI builder, along with the new Claw 3.5 Sonnet model to write and run code in an E2B sandbox. We'll go over the basics, like setting up E2B, writing our first code,

[01:03] and eventually generating graphs by uploading files directly to the chat. You do not need any coding experience to follow along with this video. Just know that this solution allows our AI agent to write code behind the scenes in order to produce outputs like charts and graphs. If this is your first time using Flowwise, then I highly recommend checking up my beginner series over here. For this video, I'll just run Flowwise on my own machine. Let's start by creating a new chat flow. Let's give it a name, like data analyzer,

[01:33] and let's save this. Now let's create our agent. Let's go to add nodes. Then under agents, let's add the tool agent. The tool agent node allows us to create AI chat paths that are able to use tools, like the E2B code interpreter tool. Let's start by adding our LLM. Let's go to add nodes. Then under chat models, I'll be using the new Claw 3.5 Sonnet model in this video, so I'll add the chat anthropic node to the canvas. Under connect credentials, click on this

[02:03] dropdown and click on create new. Let's give it a name, like youtube tutorial, and for the API key, we need to generate a key from the anthropic website. Go to anthropic.com, then under this clawed menu, go to API. And from this page, click on start building. You will be asked to create an account, and anthropic should assign some free credits to your account. From this dashboard, click on get API keys, then click on create key,

[02:34] give it a name, like youtube tutorial, and select the default workspace. Then copy your key, and add it to flow wise. And let's click add. And please use your own key as I will delete this key after this recording. Now under model name, let's select claw 3.5 Sonnet latest. Then let's change the temperature to a lower value, like 0.2. Now let's connect our chat model to our agent, like so. Let's also assign memory to this agent,

[03:04] and memory will simply allow us to have follow up conversations with the agent. Let's go to add nodes, let's go to memory, and within memory, let's add the buffer memory node. And then let's connect our buffer memory node to our tool agent. So it should look something like this. Lastly, let's assign a tool. Let's go to add nodes, then under tools, let's add the e2b code interpreter. And let's actually connect this to our agent as well. This code interpreter node also requires credentials. So click on this

[03:35] drop down, click on create new, give it a name, like e2bapi. And now we need to create an e2bapi key. Go to e2b.dev, then click on start for free, then sign up for an account, then click on keys, click on add API key, then copy the key, and back in flow wise, simply paste in that key, and click on add. Now we're nearly done with this. Under the tool agent node, click on additional parameters. And let's

[04:08] change the system message to something like, use the provided code interpreter to execute Python code written by yourself and return the response back to the user. Now let's save this chat flow. And we should now have an AI agent that is able to write and execute code using the code interpreter by e2b. Let's test this out in the chat by asking it to write very simple Python code. For example, let's do something like write some simple code to print Hello World. If

[04:38] we set up everything correctly, then you should see a response similar to this. The AI model is simply telling us what the response was from executing the code. And if we look at the code interpreter tool, we can see the code that was generated by our LLM, which in this case is a very simple print statement. And the output of that tool is the text Hello World. And this is typically something that you would see in the terminal after executing this Python code. To demonstrate this even further, let's try a simply more

[05:09] complex example, like how many R's are in the word strawberry. Let's run this. And our agent is saying that there are three R's in the word strawberry. And looking at the code interpreter, we can see the Python code that was written in order to count these letters, as well as the output from executing this code. So let's take this one step further. Let's see if our assistant can generate graphs. Let's do a simple example like demonstrate the concept of supply and demand using a graph. And

[05:40] let's see what we got back. We can see the graph as well as an explanation of the economic concept of supply and demand. We can also see that the code interpreter was indeed called along with the Python code that was generated. And we might not see anything in the tool output. But this artifact was indeed returned. Now let's take this even further. We want the ability to upload files from our machines in this chat window, and then ask the agent questions and follow up questions based on that data. So for this tutorial, I've prepared the

[06:12] following file, which you can download from the description of this video. And as an added bonus, you can also download my workflow from the description as well, absolutely free. So if you want to show your appreciation, then please hit the like button and subscribe to my channel. This file simply contains information about 20 companies, like the company name, the ticker, the stock price, the profits, loss, etc. So back in flow-wise, let's enable the ability to upload files in the chat by going to

[06:42] settings, configuration, let's click on file upload, and let's enable file uploads. Let's save this. And now when we open the chat, we are indeed able to select files. So I'm going to clear the chat, then I'm going to select that file from my machine. And let's start with a very simple question, like describe the contents of this file. And looking at the response, we can see that the file structure is definitely correct. And there are some other interesting information about the contents as well. So this looks good.

[07:13] Now let's ask our assistant to generate a chart for us. So let's ask it something like, please generate a stock price versus gross margin scatter plot. Right, let's have a look at the response. And look at that, we do get our scatter plot back. Let's try one more example, like please also create a horizontal bar chart that shows the profit and loss. So I think you can already see the value in using something like this to create educational agents and data analyst agents as well. Let's see what we got

[07:43] back. And this is perfect. So if you are curious about the type of things that you can do with e2b, then I highly recommend checking out the code interpreting section in the documentation. Like if you go to analyze data with AI, you can see all the pre installed libraries that you can use. So if you are familiar with using Python, then this list of libraries might be important to you. If you enjoyed this video, then please hit the like button and subscribe. And if you want to see more Flow-wise related content,

[08:14] then check out this video over here. I'll see you in the next one. Bye bye.

Create an AI Agent that Writes & Runs Code in FlowiseAI (No Coding Required) ​

Build Process ​

Key Components ​

Transcript ​

Create an AI Agent that Writes & Runs Code in FlowiseAI (No Coding Required)

Build Process

Key Components

Transcript