Welcome to Lightcast MCP Documentation

Lightcast MCP provides AI agents with structured access to Lightcast labor market & skills intelligence.

Built on a semantic framework capturing user intent, it abstracts the complexity of labor market data enabling AI agents to interpret requests and apply the right analytical capabilities to deliver workforce insights grounded in trusted Lightcast datasets.

Lightcast MCP provides a curated MCP interface, not direct access to every underlying Lightcast API. It exposes validated datasets, standardized tool contracts, and guided analytical workflows that simplify how AI agents interact with labor market data.

Available Data Domains

Job Postings

Real-time labor demand from deduplicated online job advertisements

Profiles

Labor supply signals from professional profile records

Core LMI

Government sourced labor market data

The server exposes two categories of tools:

Metadata tools - for dataset discovery and canonicalization
Analysis tools - for counts, estimates, distributions, rankings, time series, and job posting samples

Purpose

Beyond raw API access, AI agents must understand available data, validate inputs, resolve natural language into canonical taxonomies, and avoid incompatible tool-dataset combinations.

Lightcast MCP addresses these needs:

Dataset Discovery

Schema Discovery

Normalization

Dataset-aware Analytics

Input Validation

Built for Builders

AI Engineers

Agent tooling

Developers

App building

Solution Architects

Workflow design

Product Teams

Prototyping

Enterprise

MCP environments

Scenarios & Use Cases

Lightcast MCP helps agents answer questions such as

Example questions

Which employers are posting the most jobs for a role in a region?
Which skills are most common in a workforce segment?
What is the median advertised salary in a job posting market?
How has posting demand changed over time?
What are the occupational earnings percentiles in a region?

Lightcast MCP helps agents avoid common failure modes by:

Guessing unsupported metrics or dimensions
Calling a tool with the wrong dataset
Passing raw user text where a canonical taxonomy ID is required
Skipping required filters, such as the time filter for job_postings analysis tools

How MCP Differs From Direct API Access

Direct API access is appropriate when an application needs full control over the underlying Lightcast API. Lightcast MCP is optimized for AI agents that need a smaller, discoverable, schema-validated interface.

Lightcast MCP

Direct API access

Gives agents a shared MCP tool model across implemented datasets.

Requires the client to understand each API's request and response model.

Exposes normalized datasets, dimensions, metrics, filters, and tool contracts.

Exposes API-specific concepts and request shapes.

Provides classify and tool_info as first-class workflow steps.

Requires separate handling for taxonomy lookup and validation.

Exposes selected implemented capabilities only.

Gives broad API access wherever credentials allow.

Recommended Agent Workflow

Identify Dataset

Match the question to the right dataset

Pick A Tool

Choose a tool that supports the dataset

Fetch Schema

Get the input schema for the dataset + tool

Resolve IDs

Convert raw text to canonical IDs if needed

Run Analysis

Call with schema-valid inputs only