CadGenie

CAD Genie is an AI-powered civil engineering design platform built to be the next-generation alternative to AutoCAD and CATIA — tailored specifically for structural engineers, architects, and construction planners.

It leverages GPT-4, OpenCascade (OCCT), and Qt6 to provide a smart, 3D modeling space where users can design buildings using natural language, instantly generate structural layouts, estimate material quantities and pricing, and export 2D/3D-ready technical drawings — all from a modern, dark-mode CAD workspace.

🏗️ What It Does: 🔷 Civil-Specific Design: Users can prompt:

“Design a 2-storey RCC building (12m x 10m) with sloped GI roofing and isolated footing”

AI returns:

Columns, beams, slabs, roofing layout

Material requirements: steel, concrete, sand, cement, bricks

Load calculations (Dead, Live, Wind, Seismic)

Indian standard code references (IS 456, IS 800, IS 875, IS 1893)

Multiple design options (compact, economic, high-strength)

Auto-generated Bill of Materials with current market prices

🖥 CAD Workspace: Built using Qt6 + OCCT for real solid modeling

Dark-mode 3D canvas with:

Centered XYZ axis

Grid snapping

Clean geometry rendering (not just placeholders)

Exploded view toggle

Roof modeling (gable, flat, shed with slope & supports)

✏️ Tools (Top Toolbar): Column, Beam, Slab, Wall, Footing, Roof, Window, Door

2D floor plan sketch → 3D extrusion

Rotate, Trim, Mirror, Offset, Snap-to-grid

Load Estimator: Live load, wind zones, earthquake categories

💬 AI Assistant (GPT-4-Powered): Compact chat input on top-right corner

GPT-4 responds with:

Structural specs

Justification for component sizing

Code references (IS 456, IS 800, etc.)

Variants for cost optimization

Editable part tree

📚 Real Engineering Knowledge (60+ sources): Includes NPTEL courses, structural textbooks, and full IS code PDFs

Examples:

Design of Steel Structures (NPTEL)

Why Buildings Stand Up by Salvadori

IS 456, IS 800, IS 875 (Parts 1-5), IS 1893

Connected to a Vector Database pipeline for future “Ask the Code” or “Query Textbook” capabilities

📤 Export Features: 2D Drawings: DXF + PDF (plans, sections, elevations)

3D Files: STEP, STL, IGES

CSV: Bill of Materials with live pricing

PDF Design Report: Calculations + References

🧪 MVP Walkthrough: User prompt:

“Design a G+2 RCC frame building with 10m x 15m footprint and flat roof”

AI returns:

Columns: 300x300 mm @ 4.5m spacing

Beams: ISMB250, 5m span

Roof: Flat RCC slab, 150mm, M25

Footings: Isolated, 1.5x1.5m, 1.2m deep

Materials: 4.2 tons steel, 18 m³ concrete

Pricing: ₹6.8L INR

Design rendered instantly in 3D

User adjusts column size → AI recalculates beams/slabs

Export all: DXF, PDF, BoM, and calculation sheet

⚙️ Tech Stack: Frontend: Bolt (for layout), Qt6 (for GUI), OCCT (for modeling)

Backend (future): Flask + LangChain + Chroma for vector database

AI: GPT-4 with embedded knowledge from textbooks, PDFs, lectures

🌱 Status: ✅ MVP is live and working in Bolt

✅ AI design assistant integrated

✅ Canvas tools + rendering functional

🧪 Still in Beta – working to add:

Live PDF ingestion (IS codes, textbooks)

Structural simulation

Real-time team collaboration

📈 Actively being developed for full-scale deployment

📣 Call to Action / Vision: I’m currently building this solo and seeking collaborators, investors, and civil/structural domain experts to take this to production.

The vision is clear: A platform that replaces thousands of hours of manual CAD + design calculation work with one intelligent AI assistant.

Let’s reinvent civil engineering design — from blueprint to build site.

Built With

• bolt
• python