MCP server

Canadian money data for your AI agent

The same engine behind metrestick.ca, exposed over the Model Context Protocol. Ask an assistant Canadian money questions — income percentiles, take-home pay, mortgage affordability, retirement, federal benefits — answered from official Statistics Canada and CRA data, with the source vintage attached. Free, open, and stateless (no accounts, nothing stored).

Connect

Streamable-HTTP endpoint — point any MCP client at it:

https://mcp.metrestick.ca/mcp

Claude Code

claude mcp add --transport http metrestick https://mcp.metrestick.ca/mcp

Cursor / generic mcp.json

{
  "mcpServers": {
    "metrestick": { "url": "https://mcp.metrestick.ca/mcp" }
  }
}

19 tools

Benchmarks

Income percentile (individual or household)
Take-home pay (2026 federal + provincial tax, CPP/QPP, EI)
Net-worth percentile by age
Raw percentile benchmarks + data vintage

Housing

Mortgage affordability (GDS/TDS stress test)
Income needed for a home price
Rent vs. buy (break-even year)
Months to a down-payment target
Housing-cost burden (“am I house-poor?”)

Retirement

Retirement target + on-track check
After-tax retirement income (incl. OAS clawback)
OAS clawback
GIS estimate
Pension income splitting
RRSP-meltdown drawdown simulation

Family & couples

Second-income value (after tax, benefit clawbacks, childcare)
Federal family benefits (CCB, GST/grocery, CWB)

Cost & wages

Cost of living by province
Occupation wages (national + provincial)

Sources & privacy

Statistics Canada (income & household percentiles, Survey of Financial Security, Survey of Household Spending, wage tables) and CRA 2026 tax + federal benefit parameters (CCB, GST/grocery credit, CWB, OAS/GIS/CPP). Every data tool returns its source vintage so answers can be cited. The server runs stateless: it computes on the inputs in each request and stores nothing.