Requirements in creating generated 10,000 images for an NFT collection

So with all the NFT Buzz going around, I just thought of creating a collection of images for fun. This specifically follows the X number of images model (usually 10k or so), with assets changing, resulting in various 1 of 1’s, giving people unique assets. Very known collections under this category are BAYC, Cool Cats, Doodles, and DeadFellaz, to name a few.

If you still haven’t read up: Understanding NFT for dummies. Intro for artists and designers. Spoiler alert, it didn’t magically make me rich! But I thought of sharing the process here just for those who were wondering about things they should know, whether as a client or a creator.


The very basic foundation of the known NFT collection is to create multiple pieces of images unique to each other in one property asset or another. With that, you have to have an idea how to create graphic images and illustrations, and have a good sense of layering.

This means you should be well acquainted with Photoshop and/or Illustrator to make your images. You can either hire an artist or make this yourself, with the latter option making it really easy for trained designers because that gets rid of that expense.

The tricky part here is doing code. There are a dozen of guides online, but from here I’d probably recommend Hashlips’s guide, which is based on Solidity. You can check his Youtube channel for guides dependent on the type of ‘mixer’ you need, up to the smart contracts on which coin economy you plan on launching.

Disclaimer: I'm very much fluent in HTML & CSS, with experience in PHP and a certificate in C++. I'm not sure if it contributes to the level of difficulty in understanding the logic behind much of these things, for somebody who's starting with any kind of language background.


One thing you have to take note is the number of possible combinations per asset. I think the easiest way to calculate this is simply A x B x C, where the variable stands for the number of assets per property. So let’s say that’s Eyes (A), Nose (B), and Mouth (C) with 4-3-2 assets respectively. That leaves you with 24 possible combinations.

Needless to say, the more you make, the more combinations you’ll get, with some traits/assets becoming rare in the process. If you don’t like doing the math, just set a number like 10 variants per asset for 5 assets. Or you could wing it, and you’ll either find yourself short, or find that each of your generated image is quite ‘rare’ in terms of probabilities. In any case, you should always keep in mind what you want to make ‘rarer’ than the rest, as there is always a means to tweak a particular probability for an asset.

Looking at the CryptoPunks above, you’ll notice that sometimes, it’s as simple as a tweak of a beard, earring accessory, or eyewear, especially evident on the top orange-hair variety.


Here’s potentially the most financially testing part of the process. You have two (2) options here.

Manual uploading is when you create a collection on OpenSea and simply upload each and every generated image one by one. You have to manually enter each asset’s traits, make sure you don’t overlook or double in uploads. This is doable. It’ll take a couple of days for a 10,000 collection, but it certainly is doable. Take note that you don’t just manually upload the images, but you need to manually type every metadata property accurately. This isn’t a problem when you automate it, which leads us to our second option.

Smart contracts, on the other hand, is more automated. But like any other easier option, it comes at a price tag. When I tried this out, uploading a smart contract would initially cost you the equivalent of USD800 to USD1500 on the Ethereum economy. It all depends on the complexity of your contract, and how much it needs to do upon contract launch. This opens up a lot of control for the creator though. Unlike the manual option, you will have a backend function of a variety of controls such as whitelisting, reveals, and airdrops. Note that every thing you do on the smart contract entails a gas fee, so every mistake or repetitive action will cost you.

Going the smart contract route would also entail you use IPFS (Inter-Planetary File System) platforms. Simply put, this is how HTTP:// functions for the internet, only more encrypted, and that’s probably the best way to describe its purpose. You would need to upload your assets, both the images and the metadata (JSON), onto your IPFS, which will be synced with your smart contract. Addresses here are like tokens. Be sure to be correct the first time you input your details, or you’re subject to overwriting and more gas fees.


That’s the basics and general understanding you need to be aware of in creating an NFT Collection under the randomly generated image 1 of X model! I did mine, and successfully created 10,000 images, but I didn’t bother creating a smart contract launch for it, simply because I just wanted to figure out that I can do it myself. In fact, I only managed to manually upload 400 images, got tired, and called it quits. That’s enough experience. LOL. I did, however, manage to play around the probabilities, since there are ways like finding faucets for test networks and such, so I encourage you to try that first if you’re figuring out how those contracts even work. It’s honestly fun from my perspective.

Don’t forget that you always have other ecosystems to tinker with, which is far more cheaper than the usually ridiculous gas fees ETH has. On top of that, the 10,000 image-nft isn’t always the route. You don’t have to get into this business model if it’s not a fit. What you must know and realize though is the fact that this disruption breeds great innovation. It doesn’t hurt to open your mind to the potential of what you can do with this technology.

You have the opportunity to utilize this NFT craze to get support where you can’t find it, crowdfund your business ideas with a roadmap, or simply to do however you please. It goes without saying you need a good marketing and community building mechanic, because it would be useless to launch a collection if nobody was there waiting to mint or purchase your items in the first place. Have a vision. Have a mission.

Clients, if you’re looking to make a collection, just know that timeline-wise, this should command at least a month for a decent output of assets. You have to take into consideration the style, the assets, and variations, alongside the approval of what is wanted and what can be replaced. On top of that is the mixing and matching, and even rarer 1of1’s that normally come out as special edition.

Good luck with your collections and NFT uploads and ventures!
Oh and just in case you were wondering, this was the useless NFT Collection I made.

Hit enter to search or ESC to close