Physical material libraries are still one of the most useful parts of a CMF studio. We can pick something up, move it under a light, and understand qualities that aren't easy to communicate through a render. The limitation is that a physical library doesn't move very easily into the rest of the digital design process. If we want to see a textile across an entire seat or compare three plastics on an instrument panel, eventually those surfaces need a digital representation.
Digital material libraries solve part of that problem. They let us reuse materials across vehicles and visualization tools, but there's another opportunity we think is more interesting: the library can become a place where new material ideas start, not just where finished materials are stored.
A digital material library is more than a texture folder
At the simplest level, a material library can just contain reusable PBR materials. That's already useful because it prevents us from rebuilding the same plastic, textile, or leather every time we start a project. Automotive workflows can go considerably further.
Porsche has presented a digital color and material workflow managing more than 900 digital materials. Its process also includes calibration across scenes, displays, light booths, and other visualization environments. Autodesk's overview of Porsche's workflow shows how a material collection can become part of a much larger digital evaluation system.
Once a library becomes large enough, we need more than filenames. Material type, finish, physical references, project information, and intended application all become useful metadata. That starts looking much more like a material database than an asset folder.
What happens when the material doesn't exist yet?
Traditional libraries are very good at answering one question: what materials do we already have? They're less useful when we know what we want but don't have it yet.
We might have dozens of black plastics in the library but none with the particular grain and finish we're considering. We might want a recycled textile that sits somewhere between two existing samples, or a metal finish that doesn't quite match anything we've already scanned or authored. Reusable leather materials and plastic materials still help when the match is close — try saddle leather or matte automotive plastic before generating from scratch.
This is where we've been building twirl differently. The twirl PBR material library lets us browse existing materials, but the material generator lets us create the missing one.
We can generate that material rather than searching indefinitely for something close. Now the library answers two questions: what do we have, and what do we want to make?
Concept materials and measured materials can coexist
Not every material in a digital CMF library needs to serve the same purpose. Some represent something real. We have a supplier sample, approved paint, or physical textile that we're trying to reproduce accurately. Those materials need traceability and may eventually use scanned or measured data.
Others are concepts. A generated material can exist because someone wanted to see what a warm technical textile would look like on the seat. It doesn't need to pretend that a supplier already makes it. That distinction lets us use twirl aggressively during exploration without confusing a generated concept with an approved physical material.
If the concept works, it becomes a useful reference for the next phase. We can look for an existing supplier material, develop one physically, or create a more accurate digital version.
Build the library while designing
The other benefit is accumulation. If we generate a fine-grain polymer that works well, we can keep it. The next project doesn't have to start from zero.
Over time, we can build a collection around the materials we actually use: interior polymers, woven textiles, technical fabrics, leather alternatives, metals, composites, and more experimental surfaces. That's the direction we're taking twirl. Generation and the library aren't separate ideas. The materials we create become part of the collection we can come back to.
Instead of repeatedly searching the internet for textures, we gradually build a material library around our own work. For a transportation designer, that means less time rebuilding the same surface categories and more time comparing the directions that actually matter to the vehicle.
Search becomes part of the design process
Once we have hundreds or thousands of materials, finding them matters almost as much as creating them. We might want to search by material type, color, finish, project, supplier, or application. For a larger CMF organization, properties such as manufacturing process, recycled content, weight, and sustainability data can eventually matter alongside the visual asset.
Recent 2026 research involving automotive CMF professionals found that fragmented material information and access to reliable data remain real workflow problems. The researchers looked specifically at how AI systems could help designers work with material and circularity information while maintaining traceability and designer control. The Design Society paper is a useful snapshot of what CMF teams actually want from more intelligent material systems.
There's a much larger material-management problem there than twirl currently solves, but it points in a direction we care about. The library shouldn't just be somewhere materials go after they're finished. It can become part of how we find, generate, and compare material directions in the first place.