Should you use our Open Source Version?

Should you use our Open Source Version?

We're super proud of being an open source SaaS vendor. Our entire chatbot technology is 100% open source, and if you want to use our open source version, we will not stop you in any ways. In fact, we're trying our best every single day to make it even easier for you to use our open source version, something you can see by the 250+ pages of documentation we're constantly adding to every single day, and 20+ hours worth of YouTube videoes we'we created to document our platform.

We've even created Docker images to make it easier for you to deploy. However, if you chose our open source version you need to chose it for the right reasons. If you chose it because you want to save a couple of hundred bucks per month, I suspect you will quickly realise that these "savings" rapidly could become very, very, very expensive.

You're paying for what you don't see

This morning I woke up to a PM from Aria. She was making a demo for one of our clients, and she wanted to create a landing page with a full screen chatbot. As she created the demo she found a bug in the inclusion script that prevented it from working. I spent two hours fixing the bug, updated our Docker images, and upgraded the client's cloudlet. Later this fix will be available for everybody.

We didn't even tell the client!

Roughly twice every single month we will upgrade all cloudlets, add new features, and fix bugs. 90% of the time we do this, we don't tell anybody, we just do it. In addition we take backup of cloudlets, and we test each chatbot periodically to make sure it's working. The other day another client had a bug in his configuration. The client probably wasn't even aware of it, but as we ran through testing all chatbots, we found the bug and fixed it. The bug was due to something the client had done, but we fixed it ourselves, without even notifying the client.

You're not paying for software, you're paying for service. The open source "free" version doesn't come with service!

Kubernetes and infrastructure

It took us 3 months to stabilise our Kubernetes cluster. It's got load balancing, cloud 7 volumes, and it's basically the same infrastructure as Google runs on. Can you run it on a $12 droplet from DigitalOcean? Of course! Will it work "most of the time"? Probably!

However, it will not perform as good as our stuff. Simply configuring our Content Delivery Network provider took us weeks. We've got one client with 1.5 million page views per week, implying the chatbot embed script is being downloaded 1.5 million times every single week. Our contact with this client probably doesn't even know what a CDN is, or why having one is important. If they switched to hosting their chatbot DIY style, they would rapidly need to understand what a CDN is I suspect, and the importance of having one ...

For the record, we're paying $200 per month for our CDN

Configuring your chatbot

When we get a new client, we spend a lot of time studying that client's business. If the client is an E-Commerce client, we ask what E-Commerce system they're using. Then we configure direct API integration towards their E-Commerce platform, resulting in 10x quality. If it's a support chatbot, we'll investigate what ticketing system they're using, and sometimes be able to directly integrate with their existing ticket system of choice.

For us making a theme according to your existing website's color profile is a 15 minute job. For you simply reading the documentation for our themes and understand how to correctly create a CSS file and embed into the chatbot is probably a 3 hours job, while experimenting and doing 1,000 things wrong in the process.

These decisions and this process is based upon experience, a lot of experience. Can you acquire the same experience? Probably, but it will take you months to acquire. Unless your time is worthless, and you've got nothing better to do, the math is quite simple. Basically ...

Using our open source version to save money is madness!

For the record, some of our competitors whom we like to compare ourselves with are charging €1,495 and €2,995 per month for their plans. We're charging $300 and $800.

Who should use our open source version then?

It's really quite simple, our open source version is there to "serve the underserved". With underserved we mean people and companies who cannot afford to pay for services. Examples are students, early stage startups without funding, voluntary organisations with zero IT budgets, etc.

We're releasing the free version for the same reasons some grocery stores will give expired food to homeless shelters. We want to be nice to those without means!

However, if you're in the business of making money, using our open source version to save money is quite frankly madness. When that's said, if you think I'm wrong, I wish you good luck. Just don't complain to us when you've spent two months worth of software development salary, and you're still not able to provide more than 50% of the quality we would give you for $198 per month. I don't know what software development salaries you've got in your company, but in our company we're definitely paying more than $198 per month ...

We're a business, we're in it to make money. Being able to give away free chatbots to those who can't afford our services is just a "nice side effect" that makes us slightly happier than we'd be in a traditional company, making it a little bit easier to wake up in the morning, realising we're also helping those who can't afford our help.

It's our little gift to the world!

If you're in the business to make money, we suspect it's much less expensive using our services than trying to figure out how to use our free and open source version. If you don't believe me, feel free to use our free version. You could probably get it up running locally on your development machine in an hour, and deployed to a droplet from DigitalOcean in another hour.

But getting it to work, with the same quality as we're able to provide you for $198 per month, will be much more expensive than $198 per month, unless your time is worthless, and your developers are working for $98 per month ...

People come to us to reduce employee workload. If you don't have $12 to pay for a DigitalOcean droplet, you could in theory replace our AI chatbot with 50 people answering emails using your local library's internet connection and computers. I suspect this wouldn't save you anything though, but if you can somehow get 50 people to work for free, the costs would be literally zero!

One of our clients have 15 full time support technicians. For this particular client, the amount of support tickets was reduced by 30% the month they implemented our chatbot. This is $240,000 per year in salary cost savings. This client had tried for months to implement their own chatbot, and tried multiple "inexpensive vendors" before they came to us. This particular client also employed dozens of senior software developers themselves, so in theory they should be well equipped to chose a DIY solution. The salary costs alone for their little "DIY project" was probably 5 digits. With us you could have a chatbot for 3 decades for that price!

Needless to say, but they're very happy with our service today! If you're not in the business of saving money and earning a profit, then please go for our open source version - However, if you're in a business, and you've got employees, and you're looking to save money - $198 per month is literally free! We would know, because ...

We spent 7 digits creating the platform!

If you've got 5 digits to spare for a "DIY solution", or 300 hours of manpower to throw at the problem, then be our guest 😊

Thomas Hansen

Thomas Hansen I am the CTO of AINIRO.IO AS. I am a software developer with more than 25 years of experience. I write about Machine Learning, AI, and how to help organizations adopt said technologies. You can follow me on LinkedIn if you want to read more of what I write.

Published 18. Jan 2024