Ask AI about salaries
What is this?
The salary statistics on this site are also available to AI assistants through an MCP server (Model Context Protocol). Once connected, assistants such as Claude or ChatGPT can answer salary questions using the same official data from national statistics offices — no scraping, no guessing.
Server address
Point any MCP-compatible client at this URL. The endpoint is public and read-only; no account or API key is needed.
https://salarydata.eu/mcp/salaries
How to connect
Claude
Claude (web or desktop): Settings → Connectors → Add custom connector, then paste the server address.
ChatGPT
ChatGPT: Settings → Connectors → Developer mode → Create, then paste the server address as an MCP server.
Other MCP clients
Other MCP clients (Cursor, Claude Code, etc.) register the server in their configuration:
{
"mcpServers": {
"salary-data": {
"url": "https://salarydata.eu/mcp/salaries"
}
}
}
Example questions
Once connected, salary questions can be asked in plain language:
- "Where does a €2,500 gross monthly salary fall in the national distribution of Latvia?"
- "Which sector pays the most in Estonia?"
- "How large is the gender pay gap in Latvia?"
- "What is the average salary by region in Latvia?"
- "What is the average advertised salary for developers in Latvian job ads?"
Available tools
The server exposes five read-only tools backed by the same data as this site:
- salary-percentile-tool — position of a gross monthly salary in the national distribution of eight countries (LV, EE, LT, NO, FI, SE, DK, PL).
- country-overview-tool — headline statistics for one country.
- sector-stats-tool — average salary by sector (LV, EE, DK).
- region-stats-tool — average salary by region (LV, EE).
- gender-pay-gap-tool — gender pay gap time series (LV, EE).
- job-ad-stats-tool — advertised salaries by company and role (job-ad data, not official statistics).
Good to know
- Estimation precision varies by country: Latvia uses exact income brackets, while others use decile or quartile bands. Every answer states its basis and period.
- The endpoint is public, read-only and rate-limited; it never exposes personal data — all statistics are aggregate counts.
- Every response names the national statistics office behind the numbers. Data reuse is subject to each office’s published license.