construction-job-costing-agent
MCP server providing construction job costing data and interactive visualizations for a fictitious general contractor, enabling queries on projects, budgets, cost codes, change orders, and cash flow.
README
Construction Job Costing Agent — MCP Server + MCP Apps
A proof-of-concept MCP (Model Context Protocol) server that provides construction job costing data and interactive visualizations for Pacific Northwest Builders Inc., a fictitious general contractor.
What This POC Proves
- MCP as a data backend for Copilot — An MCP server can serve structured financial data (jobs, cost codes, change orders, cash flow) to M365 Copilot via SSE transport.
- MCP Apps for rich UI — Tool responses include
ui://resources with self-contained HTML views (SVG charts, interactive tables) that render inline in the Copilot panel. - Domain-specific agent — A declarative agent manifest gives Copilot the persona and instructions of a construction finance analyst.
- No external dependencies for charts — All visualizations use inline SVG — no Chart.js, D3, or other libraries.
Architecture
┌─────────────────────────────────────┐
│ M365 Copilot / Chat │
│ (Declarative Agent: Job Cost │
│ Analyst) │
└──────────────┬──────────────────────┘
│ SSE (port 3001)
▼
┌─────────────────────────────────────┐
│ MCP Server (Node.js) │
│ ┌───────────┐ ┌────────────────┐ │
│ │ tools.ts │ │ data.ts │ │
│ │ 7 tools │ │ 6 jobs │ │
│ │ │ │ cost codes │ │
│ │ Handlers │ │ change orders│ │
│ │ + UI refs │ │ cash flow │ │
│ └───────────┘ └────────────────┘ │
│ ┌──────────────────────────────┐ │
│ │ views/ (HTML+SVG) │ │
│ │ • job-dashboard.html │ │
│ │ • cost-breakdown.html │ │
│ │ • budget-vs-actual.html │ │
│ │ • cashflow-forecast.html │ │
│ │ • change-orders.html │ │
│ └──────────────────────────────┘ │
└─────────────────────────────────────┘
Setup
# Install dependencies
npm install
# Build TypeScript
npm run build
# Start the server
npm start
The server starts on http://localhost:3001 with SSE endpoint at /sse.
Tools
| Tool | Description | UI |
|---|---|---|
query_jobs |
List all projects with summary stats | Dashboard with sparkline bars |
get_job_detail |
Full detail for one job (cost codes, subs, COs) | — |
get_cost_breakdown |
Cost by category (Labor/Materials/Equipment/Subs/Overhead) | Donut chart |
get_cost_code_detail |
Budget vs actual by CSI cost code | Horizontal bar chart |
compare_budget_vs_actual |
Cross-job budget vs actual comparison | Grouped bar chart |
get_cashflow_forecast |
Monthly planned vs actual vs forecast | Line chart |
get_change_orders |
Change orders with status filtering | Interactive table |
Register as M365 Copilot Declarative Agent
- Start the MCP server (
npm start) - In the M365 Copilot admin or Teams Developer Portal, create a new declarative agent
- Point it to the
declarative-agent.jsonmanifest in this directory - The manifest references
http://localhost:3001/sseas the MCP server endpoint - For production, deploy the server and update the URL accordingly
Testing in Copilot
Once registered, try these prompts:
- "Show me a dashboard of all active projects" → Job dashboard with cards and budget bars
- "Which jobs are over budget?" → Filtered list highlighting Cascade Office Tower
- "Show the cost breakdown for the Downtown Transit Hub" → Donut chart of cost categories
- "What's the cash flow forecast for Cascade Office Tower?" → Line chart with planned/actual/forecast
- "List all pending change orders across projects" → Interactive table filtered to Pending
- "Compare budget vs actual for all projects" → Grouped bar chart
- "Give me the cost code detail for the bridge retrofit" → Budget vs actual bars by CSI code
Sample Data
The server includes realistic data for 6 construction projects totaling ~$113.6M in contract value:
| Project | Contract Value | Status | % Complete |
|---|---|---|---|
| Cascade Office Tower | $12.5M | In Progress | 67% |
| Riverfront Mixed-Use Complex | $28.3M | In Progress | 35% |
| Highway 101 Bridge Retrofit | $8.7M | In Progress | 82% |
| Summit Ridge Condos Phase 2 | $15.1M | In Progress | 12% |
| Downtown Transit Hub | $45.2M | In Progress | 55% |
| Lakewood Elementary Renovation | $3.8M | In Progress | 95% |
12 change orders, 8 subcontractors, monthly cash flow data, and CSI-format cost codes per job.
License
Internal proof-of-concept — not for production use.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.