Wraps the Bureau of Labor Statistics v2 API with a focus on resolving cryptic SeriesIDs before you fetch data. You get natural language search against the full LABSTAT catalog to turn "unemployment rate" into LNS14000000, then pull time series for CPI, CES, JOLTS, wages, and 20+ other BLS surveys. Handles batching (50 series per request), year windows, and server-side calculations when the survey supports them. Results that exceed the context budget spill to DataCanvas dataframes, which you can query with SQL, join across surveys, or persist for later analysis. The search runs offline against bundled flat files, so it doesn't burn your 500-query daily quota. Reach for this when you need authoritative US labor data without decoding positional survey codes by hand.
claude mcp add --transport stdio cyanheads-bls-mcp-server uvx bls-mcp-server