Wraps the US Bureau of Labor Statistics v2 API with seven tools that handle the full workflow from series discovery to analysis. Start with bls_search_series to decode natural language queries into cryptic SeriesIDs like LNS14000000, then pull time series data for up to 50 series at once with optional period-over-period calculations. Large result sets automatically spill to a DuckDB-backed DataCanvas layer where you can run SQL queries across multiple series without burning through your 500 daily API quota. The search runs entirely offline against bundled LABSTAT flat files. Useful when you need CPI, unemployment, JOLTS, or wage data programmatically and want to avoid manually parsing BLS identifiers or stitching together multi-series comparisons in spreadsheets.
claude mcp add --transport stdio cyanheads-bls-labor-mcp-server uvx bls-labor-mcp-server