Adnan Chaumette, the CEO of Polygonflow, discussed GraphN, a tool that gives non-programmers the ability to make tools, with zero coding knowledge required.
80.lv: Please introduce yourself to those who still don’t know you. How have you been doing lately?
Adnan Chaumette, CEO & Founder of Polygonflow: Adnan Chaumette here, and doing great! My path into the industry has been quite unusual; I'm from Mauritania, Africa, and joining the computer graphics industry always seemed out of reach. I spent a couple of years doing 3D art, then trying to understand the tools that make 3D art, and that's how I became a Technical Artist.
Where it all began:
I was lucky enough to join Quixel, where I got to work with their brilliant team on everything, from coding to marketing, which helped me gain a lot of experience in a short timeframe.
80.lv: Let’s talk about the launch of Polygonflow. When and where did you start this path? What inspired you?
Adnan Chaumette: Between 2016 and 2017, I was releasing 3D plugins for Autodesk Maya, which helped artists in all sorts of tasks they do every day. What had started as a small project turned into a 30 thousand users business.
This is where the idea around Polygonflow started: What if I bundled all my plugins into one library, and started a basic subscription service where people would get new tools all the time?
First prototype of Polygonflow, which was the name of our library. We've changed the name to GraphN Library since then:
The idea here was clear from the beginning, and the next challenge was to actually start making every single tool. I've never been a fan of coding, and I really liked Substance Designer's approach to creating textures, so I naturally gravitated towards using Houdini to make my tools, then publishing them in Maya and other 3D software via Houdini Engine.
After playing with the software for about a week, I was back at coding again with Expressions and Python, and that's where I decided to just start creating my own visual programming software GraphN. The original idea was just to make it for me, and build tools with it, then push them into our plugins library, which was called Polygonflow at the time.
First prototype of GraphN, which basically had the looks but nothing else:
GraphN 0.8 – Early Access:
I was working at Quixel from 2017 to 2020, and when that company got acquired by Epic Games in November 2019, I felt a calling again towards this project, and together with some friends at Quixel, we had started the Polygonflow adventure in June 2020.
80.lv: How big is your team now? You shared a number of posts on hiring talented developers. How did you find the right people? What was the process? Also, how is your team organized now?
Adnan Chaumette: When the company had started, we were about 5 people, and knew we needed to expand in order to make our vision happen within the 12-18 months.
Initially, we tried to hire through various social media platforms, but most of our candidates ended up being contacts of other developers that were on the team. This was a long process that took almost a year before the team felt truly "complete", and we're still trying to fill a few key roles.
The processing of hiring, onboarding, and maintaining a team of this size is never easy, and as the hiring grows, firing becomes something that can happen. This is one of the hardest things to deal with for everyone involved, but it's an important step in the growth of any company.
Polygonflow is now an amazing team of 18 people from all around the world. Our HQ is Sweden, and we have a strong base in Hungary, which happens to host some of the most impressive programmers I've ever known. We're also fortunate to have an amazing investor who's closely involved with the project and a really good friend of mine. We have an HR Director and an Operations Officer who help us deal with all hiring, paperwork, and other tasks.
We've only recently decided to split the company into multiple "units" working to solve various problems, which is something that I should have done much sooner, given how great the results have been so far. I've been incredibly fortunate to have Ilgar Lunin as the company's CTO, and can confidently say that we'd be nowhere near our goals if he wasn't in the picture.
It goes without saying, but working with these people is an absolute blast, and I've never felt this lucky or happy in my life.
80.lv: So GraphN is an ecosystem of a toolmaker and a library; how does this ecosystem work, and how can artists use it with their 3D software?
Adnan Chaumette: Earlier, I mentioned that GraphN came to be because I wanted to make tools and improve my workflow without having to dive too much into programming. GraphN's purpose is exactly that: giving non-programmers the ability to make tools, with zero coding knowledge required.
In this video, what you're seeing is an interactive tool that creates a bridge. The user starts by creating a few objects (the green cones) that will act as the main path of their bridge, then they create a line or curve along that bridge, and from there we provide a wide range of nodes to place objects along that line, on the sides of the line, etc.
But as you can see, the user is not even touching the graph in this showcase, they're just adjusting their path at runtime and seeing it update, then adjusting the parameters in that small UI panel, and seeing their bridge update accordingly.
This is where the magic of GraphN Library comes into play: we've created a dozen of tools just like that for artists, where they don't have to interact with any node at all; they just open their 3D software, type "bridge" in the GraphN Library's search bar, then select the tool and get going.
We've barely scratched the surface with GraphN Library, and long term, we want it to host the absolute best tools any 3D artist could dream of, and when they don't find what they're looking for, there will be a thriving marketplace of tools from other artists and technical artists.
The software is standalone, yet has a deep connection into Autodesk Maya and Unreal Engine 5, with Blender, UE4, 3DS Max, and (possibly) Unity support coming down the line.
But what makes GraphN really shine is its range of possibilities: the tool in the video above is a simple bridge, but you could also make a tool that renames folders, a tool that scatters grass on terrain, a tool that does 3D Booleans or so many other ideas. This flexibility was the biggest challenge to get right, and as we started our early access, is still something that we're constantly refining.
80.lv: The toolkit now supports Maya and UE5. What are your plans for other integrations?
Adnan Chaumette: We've released our early access with support for Autodesk Maya 2020 and 2022, and Unreal Engine 5. We've got older Maya versions (2017-2019) in the works to help studios easily integrate with our ecosystem, and UE4 as well. Then we'll follow up with Blender.
For Blender, we don't want to build a full-blown integration but target specific pain points of the software like faster FBX import/export, faster Blender → UE5 data transfer, etc.
At its core, GraphN is flexible enough that we can adapt it to any software that has an extensive Python API. Maya, 3DS Max, Unreal Engine (to some extent), and Blender all fall into this category, then you have outliers like Unity and ZBrush.
Unity does have a Python API, but it's not used nor maintained enough for us to jump into it just yet, and ZBrush only has ZScripting as its API, which makes it nearly impossible to extend to the extent that we'd like to.
Working with the Community
80.lv: How do you plan to work with your community. GraphN can become a beast if users share their tools and make their own library. Do you plan to encourage sharing?
Adnan Chaumette: Absolutely! We want to see a community thriving around this ecosystem, where artists can use the tools that we provide in GraphN Library, or buy tools built by other artists and technical artists in GraphN. This will create a really healthy relationship between art and tech where it'll be easy for innovative solutions to thrive. Blender, 3ds Max, and Maya are known to have a lot of really amazing scripts that change artists' workflow for good, and our mission is to ensure that everyone can make scripts like that.
Tools like this vertex coloring system can be very niche, and only a few users would benefit from them, yet they can save weeks worth of work when used at scale. Today, we're the ones trying to find such niche but useful tools; tomorrow, the best artists on these niche topics will be making and sharing/selling such tools with the art community.
80.lv: Could you share some thoughts on the business side of things? How difficult is it to promote the new product? What are the main challenges? How difficult was it to find the right price? Also, did you consider the possibility of joining a big family like Epic or Unity at some point or would you like to keep your project independent and help all artists equally?
Adnan Chaumette: Pricing in the computer graphics software is never easy because the userbase is used to either extremely cheap or free software, in large parts because the people who build 3D software tend to have external sources of revenue (Autodesk with CAD software, Epic with Fortnite, etc...), and don't need to make their software profitable in order to survive.
For us, our ideal user is someone that may not care about making tools with nodes or having a library of tools; instead, this person would only care about one specific tool in GraphN Library, and the value of this tool will be enough for them to subscribe.
We've set our price for the Early Access to $8/month for individuals, and plan to stay within that $5-$15 range pretty much forever, as that's how much our users would be willing to spend in order to use our tech.
When it comes to potential acquisitions, I'm 50/50 on it, to be honest, but it's not something I'm particularly against or for. I think Epic, Autodesk, Adobe, Unity, and Blender all want what's best for the users because 3D software are so not profitable that you really have to do great in order to succeed, and so their shareholders' interests also align with the users.
This wasn't always the case, and there are still some decisions made here and there that may sound incredibly stupid, but at the end of the day, it's up to us individuals to vote with our wallets when that happens.
80.lv: What are your plans now, Adnan? What other new features and tools would you like to add in the future? Perhaps you would like to take a needed rest for a while personally?
No such thing as a vacation. I've been living, dreaming, and working on this project for so long, and I still feel like we're just getting started. Our team has amazing plans to transform GraphN into an ecosystem where you can build a world of tools or assets without relying on any other software, and ultimately, that's what we hope to achieve. When that's done, I hope to take some vacation.