San Francisco-based Backflip AI has made its AI copilot for CAD generally available, releasing it as an add-in for Autodesk Fusion and as a standalone web application. 

The company’s second-generation foundation model converts 3D scans, STL files and other mesh data into fully editable parametric CAD models with complete feature trees, cutting the cost of digitizing a physical part from roughly $1,500 to about $10.  CEO Greg Mark puts the turnaround at one to five minutes, against hours or days of skilled reverse-engineering labour. The technology is already in production at an unnamed top automotive manufacturer.

A model that builds like an engineer, then checks its own work

Backflip‘s model was trained to construct geometry by chaining CAD operations, extrude, revolve, pattern, chamfer, fillet, in the order a human engineer would use them, so the resulting feature tree follows real modelling conventions rather than producing a static surface wrap. 

Agent loops sit on top: the 3D Scan to CAD tool reconstructs a mesh, evaluates its own output, then iterates on fidelity and dimensional accuracy. Fast mode handles typical geometries in a couple of minutes, while Thinking mode allocates more compute to complex parts. The tool is optimized for moderate-complexity 3-axis CNC milled and turned components. A second feature, Multimodal AI Search, queries a parts library by text or photograph, a phone snapshot of a sketch can surface a matching bracket already buried in a company’s CAD archive.  

The web app exports portable .STEP files for any mechanical CAD package, and Backflip says plug-ins for further CAD platforms arrive later this year. Free accounts include four reconstructions; paid tiers start at $20 per month across Builder, Pro, Business and Enterprise plans.

Kairo Morton, Founding Engineer at Backflip, emphasized that realizing the technology’s full potential meant reimagining its underlying architecture. “Most frontier models have limited 3D comprehension that emerges more or less as a byproduct of their training on large-scale internet data. We built ours with 3D CAD operations as a first-class primitive and trained it to design the way engineers actually do – breaking down a bulk geometry, then rebuilding it feature by feature, operation by operation. The result is the first foundation model whose output you can open in your CAD package and just keep working.”

AI copilot for CAD. Image via Backflip AI.

From mesh generator to model that builds a feature tree

Backflip’s product line has moved through three distinct stages in under two years. Its December 2024 debut, “Idea to Mesh,” launched from stealth alongside a $30 million Series A and a tool that turned text and photo prompts into 3D printable parts, collapsing designs that would have taken days into minutes. 

The March 2025 foundation model reframed the problem around scan data and factory downtime, targeting digital twins that could be CNC machined or 3D printed as on-demand spares. It was trained on a synthetic dataset of more than 100 million unique 3D geometries and shipped as two tools: a SOLIDWORKS sidebar plug-in producing native, fully-featured parts with a customizable feature history, and a web app exporting STEP files. 

Mark framed the technical task then as cleanup, 3D scanners generate millions of data points but also micro surface textures that “can’t be manufactured with traditional tools,” which Backflip’s model converted into clean, manufacturable geometry. The company claimed downtime could be halved. 

The recent upgrade is a shift from cleaning up a captured surface to planning how the part gets built, and from a one-shot conversion to an agent that iterates and checks itself. 

A 3D printed copper mug designed in Backflip. Image via Backflip.

The digitization gap nobody could afford to close

Backflip’s strategy is to attack the reverse-engineering bottleneck from the cost side rather than the capture side. Scanning has been solved for years and can be taught in days; the expensive step is the conversion, which has stayed locked behind years of CAD experience and, by the company’s own figure, roughly $1,500 per part. 

The same gap has drawn very different answers across the industry. Hexagon’s Portable Metrology Division paired its ATLASCAN Pro handheld scanner with Geomagic Design X in July for aviation and defense MRO teams replacing cracked or discontinued interior components, claiming CAD rebuilds up to eight times faster than mainstream tools, with an Accuracy Analyzer validating surfaces against scan data at each stage and LiveTransfer pushing finished models into SOLIDWORKS, Siemens NX, Autodesk Inventor and PTC Creo. 

Others skip scanning entirely and mine the paperwork instead. 3YOURMIND’s Technical Drawing Analysis, added to its Part Identification software, reads legacy 2D drawings up to 200 times faster than manual interpretation and returns manufacturability assessments, cost and lead-time estimates in batches of thousands, aimed at defense, aerospace and energy operators whose only design record is a hardcopy. French start-up Spare Parts 3D built Théia on its DigiPart platform for the same problem, combining deep learning, computer vision and OCR to turn technical drawings into printable models and cutting a task that took an engineer up to a day per part down to minutes.

Everyone agrees the bottleneck is conversion, not capture. Hexagon speeds up the engineer; 3YOURMIND and Spare Parts 3D skip the part entirely. Backflip keeps the human at both ends but drops the years of CAD experience in between. At roughly $10 a part, that gets tested rather than argued.

3D Printing Industry is inviting speakers for its 2026 Additive Manufacturing Applications (AMA) series, covering Energy, Healthcare, Automotive and Mobility, Aerospace, Space and Defense, and Software. Each online event focuses on real production deployments, qualification, and supply chain integration. Practitioners interested in contributing can complete the call for speakers form here.

To stay up to date with the latest 3D printing news, don’t forget to subscribe to the 3D Printing Industry newsletter or follow us on LinkedIn.

Explore the full Future of 3D Printing and Executive Survey series from 3D Printing Industry, featuring perspectives from CEOs, engineers, and industry leaders on the industrialization of additive manufacturing, 3D printing industry trends 2026, qualification, supply chains, and additive manufacturing industry analysis.

Featured image shows AI copilot for CAD. Image via Backflip AI.