SmartSpec

Inspiration

In a prior job I spent countless hours calling manufacturers for quotes on pressure vessels. I saw firsthand how slow and inconsistent the quoting process was, some took days or even weeks. After talking to hundreds of fabricators, it became clear this inefficiency wasn’t just an inconvenience, it was a major pain point holding the industry back. That experience inspired me to build SmartSpec, a tool that lets manufacturers quote faster while keeping their focus where it belongs: on quality.

What it does

SmartSpec allows manufacturers to upload their manufacturing drawing sets, automatically extract key dimensions and materials, and generate real-time quotes powered by live commodity pricing and learned fabrication data. The system will learn from previous quotes to continually improve accuracy, cutting quote times from days to minutes.

How I built it

Built with Supabase, React, and Anthropic’s AI models, SmartSpec connects manufacturer inputs to a real-time database of materials and labor metrics. I integrated a contractor drawing dataset as a first version (since CAD extraction pipelines are still being optimized). The AI model is trained to interpret these drawings and produce preliminary cost estimates automatically.

Challenges I ran into

Connecting Supabase fully and building the AI feedback loop was tough, especially with limited time to train models effectively. Parsing and analyzing PDFs was another challenge: many CAD drawings export differently, and not all data layers are easily extractable. To stay on track, I pivoted to simpler drawing sets that simulate real manufacturer quotes to test core functionality.

Accomplishments that I'm proud of

This was my first hackathon, and I’m proud that SmartSpec already gained traction, with three manufacturers interested in piloting the system. I also learned to combine LLMs, databases, and design tools into a functioning prototype in just a few days.

What I learned

I learned how to integrate AI models into practical engineering workflows, manage backend infrastructure, and most importantly, how to validate an idea through real customer discovery. Talking to manufacturers guided every technical choice made.

What's next for SmartSpec

Working to fully automate CAD drawing analysis, expand material pricing database, and launch pilot with manufacturers. SmartSpec’s mission is to bring AI-driven precision to every fabrication shop, making quoting faster, smarter, and truer to what manufacturers care most about: quality and trust.

Built With

• anthropic
• lovable
• openai
• supabase