Breaking CAD provides an AI‑driven design workflow that lets users describe a product in natural language or with reference images and receive manufacturing‑ready CAD, STL, and SVG files. The system iteratively refines designs based on user feedback while automatically applying material tolerances, clearance rules, and stock size constraints, enabling rapid prototyping for laser cutting, CNC routing, and 3D printing. It targets product designers, engineers, small‑batch manufacturers, and makers who need fast, low‑skill generation of manufacturable designs.
Funding
Funding not disclosed
Founders
Product
Problem
Designing physical products often requires specialized CAD skills and iterative manual translation of concepts into manufacturable files, creating bottlenecks for engineers, makers, and small manufacturers who lack extensive design expertise.
Solution
Breaking CAD offers an AI-native design workflow that lets users describe a desired object in natural language or provide reference images, then iteratively refines the concept until the design meets the specified requirements. The system generates production‑ready CAD, STL, and SVG files while automatically applying material tolerances, clearance rules, and stock size constraints. Its underlying large‑language model combines natural language understanding with spatial reasoning, enabling it to interpret sketches, enforce fabrication constraints, and output files ready for laser cutting, CNC routing, or 3D printing. Users can preview intermediate results at any stage, reducing the need for manual re‑modeling and accelerating the path from idea to physical prototype.
Target Audience
Primary users are product designers, mechanical engineers, and small‑batch manufacturers who need rapid, low‑skill generation of manufacturable designs, as well as hobbyists and makers seeking AI assistance to bridge concept and fabrication.
Features
- Conversational interface for specifying design intent via text or reference images
- Iterative refinement loop that updates the model based on user feedback
- Automatic generation of CAD, STL, and SVG files compatible with common manufacturing equipment
- Built‑in awareness of material tolerances, clearance requirements, and stock dimensions
- Support for multi‑material outputs suitable for laser cutting, CNC routing, and 3D printing
- Novel LLM architecture optimized for spatial reasoning and constraint handling
- Training on curated prompt‑to‑CAD datasets combined with real engineering specifications