Backflip AI is targeting one of the slowest parts of digital manufacturing: turning existing physical components into useful CAD files. Its second AI model for CAD generation converts 3D scans and mesh files into fully editable, parametric CAD models, a workflow the company says can now take minutes instead of hours.
Why scan-to-CAD matters
Many parts used in factories still do not have usable digital models. According to CEO Greg Mark, most factories have digital models for less than one percent of their parts. That gap creates friction when teams need to prototype, update, compare, or design around existing components.
The issue is especially important in automotive and aerospace work, where digitizing components plays a key role in prototyping and designing new parts. In those environments, a digital model is not just a visual reference. It becomes part of the engineering process.
Backflip AI’s new model is built for that problem. It takes 3D scans and mesh files, then produces CAD models that can be changed after generation. The promise is not only faster conversion, but a more useful result for people who need to keep working inside CAD software.
What Backflip AI’s model produces
The main distinction is the type of output. Earlier AI models generated triangle-based surface meshes. Those can describe the outer shape of an object, but Backflip AI says its system produces actual CAD operations, including extrusions and revolutions.
That matters because the result remains editable. A parametric CAD model is meant to be modified through its underlying construction steps, rather than treated as a static surface. In practical terms, the output is designed to behave more like a model an engineer can continue adjusting.
The source material can be a 3D scan or a mesh file. Backflip AI’s model then reconstructs that input into a CAD representation with operations that can be edited. The article does not describe every part of the workflow, but the core point is clear: the tool is aimed at reducing the manual effort normally required to recreate scanned objects in CAD.
Autodesk Fusion is the first workflow
The tool runs as an add-in for Autodesk Fusion. That places the scan-to-CAD process inside a familiar CAD environment rather than making it a separate, disconnected step.
Backflip AI is offering four reconstructions for free. After that, paid plans start at $20 per month. The pricing structure gives users a small number of conversions before they need to decide whether the workflow is useful enough to pay for.
The company has been active since December 2024 and is backed by $30 million in funding. Its second AI model extends its focus on CAD generation into a specific use case: converting existing geometry into editable engineering models.
What changes for teams with missing models
The biggest implication is time. Rebuilding a part from a scan or mesh file normally requires deep CAD expertise and a significant amount of manual work. Backflip AI says its model reduces that process to a few minutes.
That does not mean every factory problem is solved by automatic reconstruction. But if the output is editable and parametric, it can be much more useful than a surface mesh alone. Teams can treat the generated model as a starting point for prototyping, modification, and new part design.
The tool also addresses a practical bottleneck: physical parts often outnumber available digital models. If most factories have digital models for less than one percent of their parts, then even a partial improvement in scan-to-CAD conversion could make a large difference to how quickly teams bring existing components into digital workflows.
The bigger CAD question
Backflip AI’s release points to a broader shift in CAD generation. The value is not just making a shape appear on screen. The value is creating geometry that remains editable inside the engineering tools people already use.
For automotive and aerospace teams, that distinction is important. A mesh can help represent a part. A parametric CAD model can support the next design decision.
Backflip AI is positioning its new model around that difference. By converting 3D scans and mesh files into editable CAD operations such as extrusions and revolutions, it is aiming to make digitized parts easier to revise, reuse, and bring into product development workflows.