CADMAGEDDON: High-Fidelity CAD Reconstruction with Local Inference
Abstract
CADMAGEDDON is an open-source agent that reconstructs editable CAD from meshes locally on single-GPU systems and laptops, using a mid-sized Qwen model and geometric tools to outperform prior systems.
CAD practitioners often work with proprietary parts that they are unwilling to share with external services. For individuals and small companies, bringing this work in-house also requires computational resources they can afford. We introduce CADMAGEDDON, an open-source agent that reconstructs editable CAD from meshes and can run locally on single-GPU systems and current-generation laptops. The method decomposes this complex task into a chain of simple visual decisions about what to construct and what to correct. Geometric tools support precise construction, allowing the agent to focus on visual reasoning. Using the mid-sized Qwen3.8-27B model, CADMAGEDDON outperforms prior approaches, including systems based on frontier models, in reconstruction quality on all the main benchmarks. CADMAGEDDON integrates with FreeCAD, bringing its reconstructions into practitioners' existing CAD workflows. Code: https://github.com/kulibinai/cadmageddon
Community
CADMAGEDDON: mesh ā editable CAD with one open-weight model, no frontier API needed š ļø
Most mesh-to-CAD systems make a model write the whole program, including every coordinate and dimension. We split each feature in two:
⢠geometric instruments detect and name the entities on the target mesh (levels, fitted lines/arcs, faces, hole stacks, patterns) and fit their exact values;
⢠a mid-sized VLM (Qwen3.8-27B) looks at renders and decides only what to build next and what to fix, referring to geometry by name.
No reconstruction model is trained. The agent emits CadQuery or native FreeCAD programs, and the model fits on a single 48 GB GPU.
Results (vs. the strongest baseline, CADFather):
⢠MCB: IoU 0.862 ā 0.933, GMS 0.705 ā 0.818
⢠CADENA-Bench: GMS 0.800 ā 0.909
⢠BenchCAD (64³ voxel IoU): 0.991, above the best self-reported frontier entry (0.963)
Get this paper in your agent:
hf papers read 2611.22563 Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper
Collections including this paper 0
No Collection including this paper