Choose the right tool¶
Start with metadata when you need to discover valid fields, models, datasets, or taxonomy versions. Use returned IDs in follow-up calls.
Skills and titles¶
| Your question | Tool |
|---|---|
| What is the standardized ID for this skill? | search_skills |
| What does this skill record contain? | get_skill_details |
| Which skills are explicitly mentioned in this text? | extract_skills_from_text |
| Which skills are related to this one? | get_related_skills |
| What is the standardized job title? | normalize_job_title |
| What titles match this phrase? | search_job_titles |
Skills search uses type IDs: ST1 for specialized skills, ST2 for common skills, and ST3 for certifications. The current search endpoint does not support category filters.
Extraction finds mentions in text. It does not tell you everything a person with that job title should know. For inferred occupational associations, use the title-to-skills workflow.
Relationships and careers¶
| Your question | Tool |
|---|---|
| Which LOT occupations fit a title and description? | classify_occupation |
| How do IDs map between taxonomies? | list_taxonomy_mappings, then map_taxonomy_concepts |
| Which skills are associated with this occupation? | query_similarity with lotspecocc-skill |
| Which occupations are similar? | query_similarity with soc, onet, or lotocc |
| What could be a next career step? | get_career_next_steps |
| What is the skill gap between two occupations? | get_career_skill_gap |
The Similarity API returns model scores and can return warnings when evidence is insufficient. Its scores are model outputs, not guarantees of an individual’s career outcome. See Lightcast’s model documentation.
Labor-market data¶
| Your question | Tool |
|---|---|
| Which US postings match these filters? | search_job_postings |
| How many matching postings exist? | get_posting_statistics |
| What are the top skills or employers? | get_posting_rankings |
| What does an occupation’s salary or demand look like? | get_occupation_benchmark |
Use get_postings_metadata to discover supported filters and metrics. Every posting query requires filters.when: either "active" or a { "start": "2026-10", "end": "2026-10" } range. Benchmark IDs must be LOT occupations or specialized occupations; SOC codes do not belong in this endpoint.
Combined workflows¶
normalize_title_and_get_skills performs normalization, LOT classification, and Similarity lookup. It deduplicates skills by ID, keeps source occupation IDs, and reports partial failures in workflow_metadata.
analyze_job_posting_skills normalizes a title and extracts explicit mentions from the title and description. Each combined skill keeps its ID and text source.
normalize_title_and_extract_skills uses Classification for both normalization and extraction. It preserves warnings and uses an active classifier release.
See the complete input schemas for all 37 tools.
Core LMI¶
Use list_core_lmi_datasets → get_core_lmi_versions → get_core_lmi_metadata → get_core_lmi_dimension → query_core_lmi to move from account access to a reproducible data query. Core LMI examples cover employment, annual earnings, regional industry analysis, and education completions, plus the observed pagination and cost-of-living limitations.