AI-Powered Job Estimation for Construction Contractors
Building a full-stack contractor management platform with AI photo estimation, receipt scanning, and real-time job costing — from zero to 47 API routes.
The Challenge
Hawaii's flooring contractors were estimating jobs with pen, paper, and gut feel. A typical bid process looked like this: drive to the site, eyeball the square footage, guess at material costs based on memory, and write up a number on the spot. Errors meant eating costs on underbids or losing jobs on overbids.
The contractor needed a tool that could: - Take photos of a room and estimate square footage using AI - Scan receipts and automatically extract line items and costs - Track jobs, materials, labor, and profitability in real time - Work on a phone at a job site with spotty connectivity - Generate professional PDF estimates to send to clients
No existing tool combined all of these. QuickBooks doesn't do photo estimation. Buildertrend is enterprise-priced. The gap was a mobile-first tool built specifically for small trade contractors.
The Solution
Claude architected and built the entire platform across multiple sessions:
AI Photo Estimation Contractors snap photos of rooms, floors, or surfaces. The app sends images to Claude's vision API, which analyzes the space and returns square footage estimates, material recommendations, and cost projections. This replaced the "eyeball and guess" workflow with data-backed estimates.
Receipt Scanning Snap a photo of a Home Depot or supplier receipt. Claude's vision API extracts every line item — product name, quantity, unit price, total — and maps them to the correct job in the system. No manual data entry.
Full-Stack Platform - 47 REST API routes covering jobs, estimates, materials, labor, clients, invoices, calendar sync, and a contractor marketplace - 20 database tables with full relational integrity - Real-time job costing: know your margin on every project as costs come in - PDF estimate and invoice generation - SMS notifications (Twilio integration) - Calendar sync for scheduling
Mobile-First Architecture Built with Capacitor for iOS deployment alongside the web app. The UI is optimized for one-handed use at a job site — large tap targets, camera-first workflows, and offline capability.
The Results
ProBuildCalc went from concept to a fully deployed platform with 47 API routes and 20 database tables:
- 4 AI-powered features: photo estimation, receipt scanning, multi-photo analysis, and smart material recommendations
- Real contractor adoption: built in direct collaboration with a Hawaii flooring contractor who uses it daily
- Complete job lifecycle: from initial estimate through invoicing and profitability reporting
- Mobile deployment: runs on iOS via Capacitor and web via any browser
- Marketplace feature: contractors can list services and find subcontractors
Every line of code — backend API, database schema, frontend UI, AI integrations, and deployment scripts — was written in collaboration with Claude across 9 development sessions.
Claude's Role
Claude designed the database schema, wrote all 47 API routes, built the AI photo estimation and receipt scanning pipelines, created the frontend UI, configured Docker deployment, and iterated on the product based on real contractor feedback. It was the sole engineering resource across 9 build sessions.
Tech Stack
Want results like these?
Let's talk about what AI can do for your business.