0:00 All right, awesome. So, hello everybody. My name is Sarah Clevinger. I am the chief product officer over at Toolkit3D. If you are familiar Oh, there we go. So, if you’re already familiar with Toolkit3D, that’s exciting. You might have been at CDFAM before. Last year we were here with Carbon where in New York City where we talked about our project with CCM where we were actually able to take their custom fit helmets from an 8-hour process to under 10 minutes from scanned to the print file, not the full print.

0:32 Of course, that still takes the time that it takes. And from here we’re able to transform their workflow from a ton of manual steps into one fully integrated automated process. So, what the heck is Toolkit3D and how does that work? We’re most known for two things. First, we are this modular platform where we take a very middleware approach where we’re integrating, we’re aggregating, and we’re automating all of your existing software stack, all of your existing software technology.

1:03 We’re not trying to get you to leave your Grasshopper workflows behind. We’re trying to automate them for you. So, we really mean it when we say modular, we’re modular to the type of scan. We’re modular to the print processes, agnostic to everything. And because it’s modular, you can take what you need and nothing that you don’t. So, if you don’t want to do scan to fit, if you don’t want to add different design engines, you don’t have to.

1:29 Our platform is really used across industries where custom fit is a big deal. So, orthotics, prosthetics, helmets like CCM. We’ve worked with vices for the NFL. If any American football fans are here, maybe not. We hate football. That’s good. All right, perfect. But if you saw a couple years ago, Mahomes had a helmet that shattered cuz it was so cold during the game. That was a custom helmet.

1:53 So, we’re able to swap his custom pads out, throw them into his new helmet, and he still had a perfect fit helmet. Really, really cool experience. And of course, one of my favorite things is footwear. So, some of my background is with HP where we launched our custom footwear projects and then I dipped out of there to do this. So, really good life choices. I’m very happy about it because the theory is that mass customization isn’t just a really cool future, right?

2:24 It’s real and we can do this now. I crossed out all of the today, tomorrow, and future because we don’t have to make very generic products anymore, right? We have the technology, we have the manufacturing capabilities, and now you have a software stack that’ll integrate everything and make it way more accessible for you. So, we’re going from that one for many to that one for one, and we’re having a lot of fun doing it.

2:51 To talk really quickly again about our workflow. This is kind of a generic experience where there’s a brand, they create a product, we have a scan, we’re taking that custom data, we’re using scan rectification and design engines, whatever those are, and then we’re throwing it over to the service bureau where we’re printing the parts and we’re doing all the post-processing. So, we’re really connecting that from end to end.

3:10 You can track every step of the way HIPACO compliant, especially when we’re in our orthotics industry. But something that I really want to talk about today conveniently circled for you is our design engine which is of course a module within the platform. And so we’ll start there. Our design engine enables these complex rule-based editing that generates the parametric 3D products and we can ingest anything from a 3D scan, a pressure reading, even prescriptions.

3:39 So, a lot of our athics and prosthetics customers just give us a PDF prescription and we turn that into something at the end, which is really fun for us. And we can do anything from a full auto design all the way to manual. So, if you want to automate it from scan to fit, absolutely, we can do that. Or you can go in and you can make the modifications yourself.

4:00 You can say, “Ah, this is a cool patient. We need to make sure that we’re addressing all of their needs.” And today, we’re introducing two new features. So first one is our algorithm as a service. This is where we’re taking all of our scan rectification that you saw from that other slide and we are pulling it out of the platform and making it a plug-in. We saw a lot of feedback that the platform is cool but we don’t need all those steps.

4:24 We really just want these little features. We’re like okay sure we can do that. We can own up to the whole module belief right. And then the second thing I’m going to show you which is super cool and we’ll have a very long five minute demo of our 2D to 3D design generation and that today is focused on footwear. So for both 3D printed or traditionally manufactured, you’ll see how we turn a 2D design into a mold that you can use for actual manufacturing.

4:53 And if I’m talking really fast, someone just wave at me. All right, cool. So, let’s talk about the algorithm as a service. And it’s not going. All right. Custom fit. It’s impossible to scale, right? You guys know this. When you’re opening your scans, you’re editing your scans, you’re cleaning them up, you’re trying to fit products to them. You hate it. It’s not fun. I get it. You need a lot of tedious hands-on work.

5:19 You need a skilled team of CAD experts and it takes hours to complete like we found out with CCM where it took them eight hours to make one helmet. Of course, we can’t scale, right? It just doesn’t make any sense. But with our new algorithm as a service in this very cool video on the side here, we can do something a little bit different. So, what that means is we’re able to take that raw scan and we’re able to rectify it, clean it up, use all of our algorithms that we have and turn any raw scan into something usable.

5:53 Our promise is any scan, any body part. And we’re already working with some companies that you really know and love like structure or snug fit. So, we’re taking their raw scans and making them usable, which is super super exciting. We can also do scans from foam boxes or from lasts. So we’re not limited to just some random foot scan or some random body scan. We have a lot of capabilities there.

6:21 But let’s talk about how it works. So first we have our proprietary modeling algorithms. We combine that with some artificial intelligence and we’re doing that reconstruction, that repair, and then all of the rectification for those scans. So you get this very messy scan and we clean it up. Step two is landmarking. We have hundreds of anatomical and non-anatomical markers where we’re able to identify what type of scan you’ve put in.

6:45 So, we know if it’s a foot, we know if it’s a hand, we know if it’s your face. We’re not just guessing there. And then we can position that scan for the proper XYZ orientation, too, so that you can put it into your software without having to make those steps. Again, it all just makes replicating those workflows that much easier. Next, we have reconstruction. Reconstruction is super fun.

7:05 We have all of the repairs that are automatic. We clean it. We again rectify it. There’s a lot of options here when it comes to your scan. So, for example, on the left side of this photo, or the right side, I suppose, you’re able to choose like to remove hair if you know that you’re working with people that have a lot of facial hair, you’re able to choose if hairstyle.

7:25 So, when we work with the NFL, a lot of those players have some crazy hairds, right? And which changes the fit of the helmet. So, we’re able to go in, we’re able to rectify what their hair looks like, so we know that the helmet’s going to fit. With CCM, for example, they have a chin strap, so we have to make sure if anyone has a beard, we’re rectifying the chin so that the chin strap fits.

7:47 A lot of really cool options here. You can also go through, you can clip shoulders, you can remove the floors for the foot and the leg scans. And then we have iterative orientation. So, if you do have a full foot scan, you’ll see this one on the next couple slides, but we’re able to go through and straighten those scans. So, if you’re building something like an AFO, you can’t build an AFO off of this scan.

8:09 It’s not going to do anything for that patient, right? So, we’re able to take that scan, rectify it, add any kind of the features that that doctor or the orthotist needs to add, and give them a product that’s going to work. And then of course there’s different levels of smoothing. So depending on what again the customer wants to do, you can make it less smooth, you can make it more smooth.

8:32 It’s really up to you. And now because talking is boring, let’s watch an actual video of the process. All right, real time. Here we go. Stay with me, guys. This is worth it. So we’re uploading the scan. This is a scan from a structure scanner. It is not a super good scan. You can say we’re really not getting a lot, right? Because we know that people are not good at scanning things.

8:54 Like we we know this. We admit to this. That’s great. So, we’re taking this barely usable scan that you would otherwise spend hours doing it and we are very quickly rectifying it to these precise specifications. This customer wanted to cut the T-zone. They wanted only the bottom of the foot. You don’t need the top of the foot cuz they’re making orthotics. And so, we’re able to do it super fast.

9:14 Another example of another terrible scan. This person scanned their foot by pulling their toe back to get that really good arch, but they didn’t get anything else. That’s okay. We don’t need anything else. And we’re able to, again, same requirements to make that T-shape, that bottom of the orthotic. This is a super cool scan, but we don’t need the full leg. So, we’re able to say, “All right, get rid of the rest of that.” Again, same requirements.

9:39 So, we’re able to take a ton of different scan data and we’re able to clean it up, rectify it, and it’s literally a web app. So, drag and drop real time. No marketing magic here. And that is the really exciting thing about algorithm as a service. So, it is that 3D scan processing. From here, we’ll let that we’ll let that go through a little bit further. Beautiful.

10:07 All right. From here within the design engine, we can go into part positioning. We can go into part fitting. You don’t necessarily have to do that. Of course, I want you to integrate with our entire platform, but again, we have that promise that we will be modular, we will be agnostic, and you guys just take the parts that you need. Now, within our design engine, we’ve taken it a little bit further with how we’re enabling the entire workflow.

10:31 So, this is just a quick render. This is printed, actually. These are pictures of a 3D printed product, but a sample of a slide that doesn’t exist for anybody that we wanted to just prove that this worked. And so what we’ve added to our design engine is 2D to 3D generation, lattice generation, all the way to mold generation. And what better example than an industry that lives and dies by their launch calendars, footwear.

11:02 So now I have a real time. Is this one playing? We’re going to make everything play. Okay. Real time example of 2D to 3D. I hope you guys can see this. Okay. Step one, you load your last. This is just a generic size 44 last. These are your lasts. Step two, Illustrator files. Super super basic Illustrator files. What we’re doing is the top line of the midsole.

11:31 We have to do it, of course, to both sides. This doesn’t have to be perfect. You don’t have to line them up exactly at the same height. You’re the footwear designer. You can get as crazy as you want. This can be your next easy shoe, but we’re doing it from 2D to 3D. Then you’re doing the bottom line of your midsole. And of course, you’re doing the outline of the entire midsole cuz we want to know how big these things are.

11:50 Again, talking really fast to keep up with this video that I did not speed up. So, you as a designer have all the freedom and all the control that you want. Now, there’s a few other things. You can’t unfortunately read them over here, but you’re able to decide what type of midsole you want, right? So, if you want something that this one is just a parabola, you have a very nice generic looking midsole, but you don’t have to stop there.

12:16 So, there’s a lot of tools right here on the side where you can choose the actual frequency of your midsole shape. So, you can add waves, you can add other features. You really have a lot of freedom within the design. I am not a footwear designer, so this is a very boring shoe, but you can do way better than me. So here we’re adding the first wave.

12:37 You can see we just did a cosine instead of the parabola, and now we have that first wave. Super super easy. It generates in seconds. We’re not This is a 5minute video. This isn’t supposed to take forever. From here, we’re going to do something extra cool because you guys are all thinking, who cares? I can do a smooth shape in Grasshopper in 2 seconds. It’s like this is not interesting.

12:57 But what we can do now is continue with this 2D to 3D experience and we’re able to emboss the actual design on the midsole itself. So exactly the same way Illustrator files, we’re using grayscale Illustrator files where you’re able to deboss or emboss based on the values of that grayscale. So more black or more white, more white, sorry. And here’s how you do that. Exactly the same way.

13:31 We are going to upload the design files. And we have guides for all of this. So if you’re thinking like that’s great, but how do I just make this up from scratch? You don’t. We have a ton of guides. We have a ton of templates for Illustrator already. You can go in, you can modify them however you want and then upload them directly to the software. I recognize I could have sped this part up, but you know, it’s really fun to watch people search for their files on their computer, right?

14:00 Right here again, you’re just seeing a 2D design. This is Illustrator scribbles. This is not meant to be difficult. So, we’ll do it to both sides. And we’ll do it to the bottom, of course. And we are adding a side for the bottom and just calling it cleats. But you can’t read that, so you’ll just have to take my word for it. And we hit generate. Super easy.

14:42 Almost. We’re almost hitting generate. I’m faster than the video today. But what’s really cool about this is that, you know, I know your footwear cycle takes about three weeks to design a shoe and we’re able to do this in five minutes. That’s pretty crazy. And can you guys see that or is it just projector problems? Everyone is hung over today. This is great. Really exciting crowd. Again, this is pretty basic designs, right?

15:26 So, what you’re able to do is go in and create your own designs. You don’t have to use my designs. You don’t have to use anything generic. We had a conversation with a footwear company that said, “That’s great, but we want to push the limit. We don’t want to do boring stuff.” I’m like, “Great. Then push the limit. Don’t let me be a footwear designer. That’s not what I want to do.

15:44 I want to make software. So, let us Oh, we’re replaying the video. We have to figure out how to switch slides. The next portion that’s also really cool is how we’re doing midsole design and we’re able to take that from the far left side, which is that midsole that we just created on the last video, but now instead of traditional manufacturing, moving it into more 3D print, the luck of the video.

16:15 There we go. More 3D print processes. So, lattising it and modifying those lattises based on function, not just fit. So again, Toolkit3D scan to fit, we love that. Let’s also do scan to function. So we have our midsole. It is exactly the same boring parabola shape midsole from the last video. And what we’re going to do is we’re going to take it, we’re going to add lattice structure.

16:42 I did not fast forward these again. I did not want the comment that was going to be like, “Oh, but you made it look so fast. It’s not really that fast.” So we’ll just sit here and let it be slow. But super fast as far as generation speeds. Now it’s already lattised and now you can go through and you can look at these lattises. You can inspect it.

17:00 You can modify any of the struts. You can make it higher density in different different areas. You can zone it out based on function. So if you know that you have a person who’s going to be wearing this shoe who is really like striking their heel versus their toe, we can modify the lattises for that function. And my favorite part here is where we actually overlay a pressure map and modify those lattises automatically.

17:55 All right, super duper easy here. Except it is the torture of watching a video. I thought this was I talked way too fast, guys. No one waved their hand at me. Sorry. Here we just generated a different shape of lattice. So you can regenerate your midsole, you can regenerate your lattises as many times as you want. There’s nothing that’s going to stop you from doing that. And of course looking at it layer by layer.

19:05 Okay, finally the part that I told you about first adding the actual pressure map over the shoe or over the midsole. We are obviously have to rotate it. We obviously have to make sure it is the correct foot, but we want to make sure that you this is a real life experience guys. We I left it all in there and quickly generate Cool. All right. Last but not least, I know I’m Duann is looking at me.

20:00 We are a young platform. We’re a young startup, right? But we’ve already had more than 300,000 products go through the platform. One of our biggest customers has been Super Feet with their custom insoles, custom fit insoles that are in retail, CCM, Vicis, a ton of really cool products. We have more than 25 other third party partners and integrations. So, if you’re working with software and you’re wondering, are we integrated?

20:23 The guess is probably yes. Statistically, we’re 48 times faster than traditional manufacturing. I love to say that. People are like, what does that mean? It’s back to the CCM example from 8 hours to 10 minutes. And we’ve gotten two industry awards so far and that’s like one a year basically. So I think that we definitely deserve applause for that. But yeah, we’re not here just to show you what we can do.

20:49 We want to learn what we’re missing. So if you saw that and you’re like, “Wow, Sarah, please never show us the full demo like that ever again.” Cool. We don’t have to do it. But there’s features that are not there that we want to keep developing. So please talk to me. Let’s make it collaborative. Let’s have a lot of fun. And let’s push the boundaries of what comes next. Thanks, guys. To learn more about the CDFAM computational design symposium series, to see the archives of previous presentations, and to learn about future events, visit CDFAM.com.