Building Dalphan AIO

What is Dalphan AIO?

Dalphan AIO is an electron-based desktop app used to automate checkout processes on different shopping platforms. The app was first inspired by other similar apps like CyberAIO and Kodai, which are both very popular in the sneakerhead community. I wanted to create a similar app since I believed I could improve upon the existing apps by adding more features and optimizing performance.

Dalphan AIO supported Shopify and Goat when it released, and was able to checkout some limited items fairly quickly thanks to the use of Turbo in the backend. The Goat checkout process was pretty interesting, as it utilized a proxy server to read the user's phone requests to Goat, and scrape akamai tokens to use across multiple other requests mimiced in Dalphan AIO. I'll likely go more in depth on this in a future blog post.

Building Dalphan AIO

When first starting Dalphan AIO, I had no idea where to start. All I knew was that I could use electron to view webpages, and that I was going to have to build a local webpage that was hosted in the window. So I did that, using only HTML, CSS, and JS. I was able to create a basic webpage that had buttons and looked like a sneaker bot.

Here was when I noticed flickering when swapping pages, and this was because electron had to actually load different html files when I clicked a button. This was not ideal, so I did some research and landed upon React. I had never used React before, and it was definitely a learning curve, but after a month of learning, I scrapped the old code and rewrote the entire app in React.

Here, I worked on the UI and UX of the app, and after a almost a year of development, I was able to release the first version of Dalphan AIO. The app could barely run a couple tasks without crashing due to some memory leaks, but it was a start. The main cause of this, was trying to remember states while swapping pages, and here is when I discovered Redux.

So I learned Redux, and scrapped the old code again, and rewrote the app in TypeScript with Redux. This was a huge improvement, and I was able to run near 100 tasks without any issues. I was able to release the app to a few friends, and they were able to run the bot without any critical issues. This was a huge milestone for me, and I was able to start working on the backend of the app.

The backend was written in Go, and was able to run tasks very quickly thanks to the use of Turbo. Building these scripts however, took a lot of time and energy as Go is a low level language, and I had to build the scripts from scratch. I was able to build a few scripts for Goat and Shopify, and was able to run them in the app. We actually had some success with the Goat scripts, however no checkouts due to low stock and not writing code to allow for instances to be run on servers.

Technical Details

The information I will share in this section is not meant to be a tutorial, but rather a high level overview of how Dalphan AIO's backend works.

The backend of Dalphan AIO used Turbo, however Turbo was only a request library. I had to architect a way to utilize Turbo and also optimize it for use in Dalphan AIO. So, I developed something I called the "Brain". The entire architecture of this system was modeled after a human nervous system, with the brain being the main controller of the system. The brain would send signals to running tasks, and these tasks would send signals back to the brain. This was helpful as it allowed tasks to spread information to other tasks, and also allowed the brain to control the tasks in a more efficient manner.

For example, if a task picked up a product earlier than any other task, it would notify the brain, and the brain would immediately send a signal to all other tasks to stop looking for that product. This was a huge optimization, as it allowed for tasks to not waste time looking for products that were already picked up by another task.