Glossary

This glossary defines the core terms used throughout Lightcast MCP documentation. Understanding these terms can help agents and developers navigate datasets, tools, schemas, and analytical workflows more effectively.

Term	Definition
Analysis tool	A tool that returns labor market data, such as counts, estimates, distributions, rankings, time series, or posting samples.
Classifiable dimension	A dimension that requires user-provided text to be converted into a canonical taxonomy ID using `classify` tool before it can be used in an analysis request.
Core LMI	Government-sourced labor market data covering occupations, unemployment, and industries.
Dataset	A collection of related labor market data that can be queried through Lightcast MCP. Implemented datasets are `job_postings`, `profiles`, and `government_occupations`.
Dimension	An analytics attribute used to filter, group, rank, or analyze data that includes occupations, skills, companies, locations, industries, education, and time periods.
Filter	A structured object that narrows the records or entities included in a tool request.
`glossary`	A metadata tool that describes available datasets, common use cases, and supported tools.
Government Occupations	The government_occupations dataset, which provides government-sourced labor market measures for occupations and geographic areas.
Government Unemployment	The government_unemployment dataset, which provides government-sourced employment, unemployment, and labor force measures for US geographic areas.
Government Industry	The government_industry dataset, which provides government-sourced labor market measures for industries and geographic areas.
Job postings	A dataset identifier `job_postings`, representing deduplicated online job advertisements and hiring demand.
MCP (Model Context Protocol)	A protocol that enables AI clients and agents to discover, understand, and interact with tools and data sources in a standardized way.
Metadata tool	A tool that provides information about available datasets, schemas, taxonomies, or supported capabilities rather than returning analytical results. Available metadata tools include `glossary`, `tool_info`, and `classify`.
Metric	A measurable value returned by a tool, such as `postings.unique_postings`, `profiles.unique_profiles`, or `lmi.median_earnings`.
Operator	A semantic rule that describes how multiple filter values are combined. Lightcast MCP supports explicit `or` and `and` operators when supported by the selected dimension.
Profiles	A dataset identifier `profiles`, representing professional profile records, qualifications, education, skills, and employer affiliation.
Rank dimension	The dimension used to group and rank results when using ranking operations such as `top`.
Streamable HTTP	The transport protocol used by the Lightcast MCP endpoint to support MCP communication and streaming responses between clients and the server.
Taxonomy	A standardized classification system used to organize and normalize concepts such as skills, occupations, industries, companies, schools, education programs, benefits, and geographic regions.
Tool Info	A metadata tool that returns the exact JSON Schema for a supported dataset and analysis tool pair.
Tool contract	The validated input and output shape for an MCP tool.