From natural language to complete PCB design in minutes, not hours.
Hardware engineers spend days—sometimes weeks—manually researching components, cross-referencing datasheets, and building Bill of Materials. What should be a creative design process becomes a tedious research task.
The pain points:
- Parsing 50+ page datasheets to verify compatibility
- Comparing pricing and availability across multiple suppliers
- Cross-referencing voltage requirements, pin configurations, and interface protocols
- Building BOMs from scratch, prone to human error
This is work that should be automated.
Jigsaw sits at the intersection of AI-powered design automation and hardware engineering. We're not replacing engineers—we're amplifying their capabilities by handling the research and validation work, so they can focus on innovation.
Jigsaw uses AI agents powered by Model Context Protocol to automate the entire PCB design workflow. Describe your circuit in natural language, and watch as our system:
- Analyzes requirements and identifies components hierarchically
- Searches suppliers for optimal parts based on specs, pricing, and availability
- Validates compatibility across all components (voltage, interfaces, pin counts)
- Generates complete BOM with real-time pricing and specifications
The magic? You see the AI reason through each component selection in real-time, just like a senior hardware engineer would.
What you get:
- Real-time component analysis with AI reasoning snippets
- Interactive PCB viewer with smart component placement
- Complete parts list with pricing and datasheet links
- Context-aware chat for iterative refinement
Modular Architecture
Self-contained MCP services handle all AI communication. The frontend streams real-time component analysis via Server-Sent Events, while the PCB viewer uses intelligent grid-based placement for optimal component organization.
Hierarchical Reasoning
AI analyzes components in logical order (MCU → Power → Sensors → Passives), ensuring compatibility at each step before proceeding. This mirrors how experienced engineers approach design.
Context-Aware Intelligence
When requirements are ambiguous, the AI asks clarifying questions. The system seamlessly pauses analysis, waits for context, then resumes—maintaining state throughout the conversation.
git clone https://github.com/yourusername/jigsaw.git
cd jigsaw/Jigsaw/frontend
npm install
npm run devSet VITE_MCP_SERVER_URL in .env for backend integration.
This project is configured for deployment on Vercel. See VERCEL_DEPLOYMENT.md for detailed instructions.
Quick deploy:
- Push to GitHub
- Import project on Vercel (auto-detects React Router v7)
- Set
VITE_MCP_SERVER_URLenvironment variable - Deploy
- Describe your circuit in natural language
- Watch AI analyze components in real-time
- Refine with chat if more context is needed
- Review BOM and export or purchase
Real-time Streaming
Server-Sent Events with intelligent buffering ensure reliable component analysis updates, even with network variability.
Smart PCB Layout
Grid-based placement system prevents overlap and optimizes component organization, creating layouts that actually make sense.
Type-Safe Architecture
Centralized type definitions ensure consistency across frontend and backend, reducing integration errors and speeding development.
- MCP API Contract - Chat API specification
- Component Analysis API - Analysis API specification
- Implementation Guide - Backend integration
Auto-routing • 3D PCB preview • Export formats (KiCad, Altium) • Direct supplier ordering • Team collaboration
Charles Muehlberger • Luke Sanborn • Phu Duong
Built at HackPrinceton 2025
Made with ⚡ by the Jigsaw Team
⭐ Star us on GitHub if you find this project helpful!
Questions? Open an issue or check out our Pitch Deck for more details.