How to Host AI Agents for $5/Month (Flowise + Hostinger Guide)
FlowiseAI v3 TutorialRecent📅 2025-06-16
Deployment Architecture
┌─────────────────────────────────────────────────────────────────────┐
│ DEPLOYMENT PIPELINE │
├─────────────────────────────────────────────────────────────────────┤
│ │
│ LOCAL DEVELOPMENT CI/CD PRODUCTION │
│ ───────────────── ──── ────────── │
│ │
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │
│ │ Code │ │ Build │ │ Docker │ │
│ │ + Test │ ──────► │ + Check │ ───► │ Server │ │
│ │ Locally │ │ Quality │ │ │ │
│ └─────────────┘ └─────────────┘ └─────────────┘ │
│ │ │ │ │
│ │ │ │ │
│ git commit auto deploy live app │
│ git push │ │
│ ▼ │
│ ┌───────────┐ │
│ │ USERS │ │
│ └───────────┘ │
│ │
└─────────────────────────────────────────────────────────────────────┘Deployment Steps
- Agent configuration
- Tool integration
- Memory management
Transcript
[00:00] So up until now we've been running flowwise locally on our own machines and you can see that in the browser as we are accessing it through local host 3000. Now this can be great if you're still learning flowwise or if you want to access local models and run everything securely on your own machine. However, this does come with a few limitations. External systems can't access your flow wise instance if it's running locally. So this means you can't really integrate it with things like WhatsApp, Telegram, or embed this
[00:30] chatbot into something like a website. For instance, customer support chat parts is actually a very popular use case. So the user could click on this little chat window and then ask questions about this business. Now, this simply won't be possible if you're running Flowwise on your own machine. External platforms like websites, Telegram or WhatsApp simply won't be able to access Flowwise if it's running on your own machine. So, if you're serious about building agents for clients or you want your agents to be accessed externally, you would have to
[01:01] deploy your agents to the cloud or to a VPS. You basically have two options. The first option and probably the easiest option for most people would be to sign up for Flowwise Cloud. You simply go to flowwiseai.com and you can click on get started to create an account. Now looking at the pricing page, Flowwise does offer a free tier which allows you to create up to two flows. This tier is perfect for learning FlowWise, but at some point you would have to upgrade with a starter
[01:32] plan which is $35 per month. Now considering that you get unlimited workflows for this price and 10,000 predictions, this is actually not a bad deal. And if you're doing this professionally for clients, $35 is actually not bad. And Flowwise will deal with the infrastructure and scaling on your behalf. Now, the alternative is to sell those flow-wise, which can definitely save you a lot of money. But keep in mind that Flowwise Cloud actually provides a lot more features
[02:02] than the community version, like the ability to work in teams and collaborate on flows. But if you don't need all of these additional features, then self-hosting flow-wise could be a very good option. So in this video, I'll show you how to self-host Flowwise on Hostinger for as little as $5 per month. And I've teamed up with Hostinger to give you an additional 10% discount on that price. In the description of this video, you'll find a link to this page on Hostinger. From here, you can simply scroll down and here we can find the
[02:32] KMV1 plan for only $5 per month. So, all you have to do is choose this plan. Then, for the period, you can select between 1 month, 12 months, and 24 months. With the 24 months plan, you'll get a greater discount, but you can also choose the 12 month plan, which is only $6 per month, or the one month plan at $8 per month. Now, I know I'll be using Flow Wise for the next 2 years. So, let's actually switch over to 24 months. Then, we don't have to change anything else on this page. Under coupon code,
[03:05] you can enter the coupon code Leon and this will give you an additional 10% off. This coupon code also works on the 12 month period. So, we still get our 10% off. And unfortunately, it doesn't work on the 1 month plan. So, I'm just going to go with the 24 month plan instead. Then, let's click continue. After completing the checkout process, you can now access your Hostinger dashboard. So, this is telling me my KMV2 setup isn't finalized yet. So, let's go to setup. Then, let's select
[03:35] our location. I'll go with United States. Then, continue. Then, for the template, let's search for FlowWise. Cool. Let's select Flowwise. Let's click on select. Let's enable this Monarch's malware scanner. It is free after all. Let's click continue. Then, let's create a root password. Be sure to keep this password in a safe place. Then let's go to continue. And finally, let's click on finish setup. This is basically creating our VPS along with the FlowWise
[04:07] installation. This can take a minute or two to complete. And that's it. We're good to go. So, let's click on manage VPS. And now all we have to do to access FlowWise is to click on manage app over here. That will open Flowwise in a new tab. And now we can set up our Flow Wise account. So, I'll go ahead and enter my details. Let's click on sign up. And there we go. We now have access to a FlowWise instance that's running on a VPS. The first thing I'm going to do is enable dark mode as light mode is simply
[04:39] blinding me. And now we can start building our very first agent. But before we do that, I do want to show you how you can keep your FlowWise instance up to date. From our Hostinger dashboard, let's simply scroll down to our different VPS's. I've got two. I've actually got N8N and Flowwise. And by the way, if you want to see how to set up the N8N instance, then check out this other video which I'll link to in the description of this video. Next to our Flow Wise VPS, let's click on manage. Then let's click on browser terminal.
[05:11] And this will open this scary looking terminal window. And no, don't worry, it's not that bad. All we have to do is run about three commands and that's all. And I'll also link to this page in the description of the video as well. And this article tells you how to update this instance. And that's all we're going to do as well. So here we have this updating flowwise section. And this is telling us to run these commands. And very important, this is saying don't worry your workflows are preserved as they are stored in a separate volume. So
[05:43] you won't lose your agents during this process. So let's simply execute those commands in this terminal. Let's enter docker compose stop. Let's press enter. And now it's saying this flow-wise container has stopped. Then we have to run docker compose rm. So let's do that. Docker compose rm. It's asking us to remove flowwise. So let's just say yes and removed. Then let's execute this command. Docker pull flowwise aai /flowwise. Then finally let's run docker
[06:16] compose up-d. So in this terminal docker compose up dash d. So if everything went well we should be able to access flowwise. So let's actually close this. And back in hostinger let's click on manage app just like we did before. And this time we didn't have to complete the signup form again which means all of our details are still there. Cool. Let's go ahead and build our first agent just to make sure that everything is indeed working. Let's go to agent flows. Let's
[06:46] click on add new. Then let's give our flow a name by clicking on save. Let's call this hostinger agent. Then let's add our agent node by going to agent. Cool. Let's connect this. Then let's double click on this agent node. Let's give it a name like customer support. Let's save this. And for the model, I'm actually going to use chat open AAI. So let's click on chat open AI. Then under parameters, let's add credentials. Let's
[07:18] call this open AAI API. Then for the API key, let's create a new key. Let's call this hosting tutorial. Let's copy the key and we'll add it to flowwise. Let's add these credentials. I'll just leave the model on GPT4 mini. We can actually close this. And for messages, let's actually assign a system message. And for the content, let's say you are a customer support. In fact, let's actually not do this manually. And let's
[07:50] use this generate instructions option and say a customer support agent for a restaurant called the Oak and Barrel. Cool. Let's click on generate. And this will now use an LLM to generate the system prompt for us. Okay. So, we get all of these instructions. So, it gives us steps, the output format, and a whole bunch of other things. I think this looks cool. Let's apply it. And I also want this agent to have access to a custom knowledge base which contains
[08:22] information about our restaurant. Let's do that. Now, I'm going to save this flow. Then, let's go to document stores. Let's create a new store and I'll call this Oak and Barrel restaurant. Let's add this. Then within this document store, let's add a new document loader. And I'm going to add two documents. The first one is a Q&A document which contains information about our restaurant like the current specials, operating hours, etc. For the text
[08:53] splitter, I'll add a recursive character text splitter. And I'll just use the default chunk size and chunk overlap. Let's click on preview. And this gives us about five chunks. Let's click on process. And let's add a second document loader. This time I'll add a CSV loader. So let's upload the menu. Let's click on preview chunks. And this gives us all the items on the menu along with their price. Let's process this. I'm going to refresh this list until everything's
[09:24] done processing. And now all we have to do is upload these chunks to a vector database. To do that, let's click on more actions, upsert all chunks. And let's start by adding embeddings. And for this, I'll use the OpenAI embeddings. So let's select our OpenAI credentials. And for the model name, let's select embeddings 3 small. And we don't have to change anything else here. For the vector store, let's select Postgress. Then under credentials, let's click on create new. Let's give it a
[09:56] name like Postgres API. And now we have to provide a username and password. Now we don't yet have our database. So let's go ahead and create it. In order to create a free Postgres database, let's go to superbase.com and let's sign into our account. Then let's create a new project. I'll simply call this Hostinger. Let's enter a password. And for the region, I'll select US East. And let's create our project. Cool. Now let's click on connect and scroll down to transaction
[10:27] pooler. Expand parameters. And now we can copy the username and add it to flowwise. Let's also enter the password that we entered earlier in superbase. Let's click on add. Then let's copy the host name and add it to flowwise. Let's copy the database name and add it as well. And let's copy the port number and add it to flowwise as well. Let's enable SSL. And then let's also add a record manager. and the record manager will
[10:58] prevent duplicate documents and remove style data for us. Let's select our Postgress credentials. Let's simply copy over the host name. Let's copy over the database name and the port number. Let's also enable SSL. And scrolling down, let's select full from the cleanup dropdown. Let's save our config. And finally, let's upsert these documents. All right. So I am getting this error message saying that self-signed certificate in certificate chain blah
[11:28] blah blah. This simply means we don't have to enable SSL. So I'll simply refresh this page. Let's disable SSL. Let's save our config and let's click on upsert. And there we go. Our documents have been upserted into the database. And we're now able to add this document store to our agent. If you ever want to add information to your knowledge base, you can simply add a new document loader. Or if you want to remove information, simply click on this dropdown and delete this record. Always
[12:00] remember to run the upsert process again whenever you change anything in this list. Right, let's go back to our agent flow. Let's select our hostinger agent. Then let's double click on this agent node. And then under this document store section, let's add our knowledge base. So let's click on plus. Let's add our document store. And in the description, we have to tell the agent when to use this document store. So let's say returns information about the oak and barrel restaurant. Cool. Now, if we
[12:33] wanted to, we could enable return source documents, and this will return the citations of this information. Now, this is something we don't typically want to show the viewers of our website. So, I'm going to keep this disabled. Then, let's save our flow. Let's test this and let's say, hey, all right, let's ask a question like, what is the price of a steak, for example? All right, so it's given us all of these options from the menu. Let's also ask something from the
[13:04] knowledge base, like, do you have a children's menu? And of course, we get a response back as well. And that's all we have to do to create this customer support agent. Now, we can share this agent with the world in many different ways. If we click on this API endpoint button, it shows us many different integration options. This embed section shows you how to embed this chatbot into websites. All we have to do is copy the script and that will show that little chat button in the website. Of course,
[13:36] we can also call this flow from Python scripts, JavaScript or any other application. We can also publicly share this chatbot by going to share chatbot. We can enable this make public option. And now anyone with this URL will be able to access this agent. So if we open this, we get this chat interface. And if we say hey, everything is still working. This is ideal for sharing this agent flow with your colleagues. So let's have a look at embedding this agent into a
[14:07] website. All we have to do is copy this code snippet and give it to your developer. However, we do live in the age of vibe coding. So what I decided to do is use lovable to generate a website for the open barrel restaurant. And now what we want to do is add our chat widget to this website. So all we have to do is say add this chat widget to the website. Then let's paste in that code snippet and press enter. And I can see this little bubble. Let's click on it.
[14:38] And how cool is that? So now you have access to an instance of FlowWise that you can access from anywhere. If you like this video, then consider hitting the like button, subscribe to my channel, and let me know in the comments what other Flowwise content you would like to see. I'll see you in the next one. Bye-bye.