Job Posting Pay Transparency Dataset: 101,432 Greenhouse and Lever Postings, August 2026

Aggregate counts derived from 101,432 open job postings collected from the public job boards of two applicant tracking systems, Greenhouse and Lever, covering 2,704 distinct companies. Snapshot taken 1 August 2026. No private data and no scraping behind a login. Individual postings are not redistributed; only aggregates are published.

Headline finding: Greenhouse's public job board API exposes no structured salary field at all (0 of 49,722 postings). Lever's API does, and 15,060 of 51,710 postings (29.1%) populate it. Yet Greenhouse employers state pay in the free-text description slightly more often than Lever employers do (37.4% against 30.3%). The willingness to disclose exists; the machine-readable field to hold it does not. Both APIs are public and this is reproducible in two requests without an account.

The dataset reports pay-disclosure rates broken down by applicant tracking system and by country, plus posting counts by seniority level, posting age and work mode.

Limitations are stated in full in the accompanying README. In summary: the pay-detection text pattern is tuned for dollar amounts and English phrasing, so every disclosure figure should be read as a lower bound, with the India, UK and Australia rows most affected. Two platforms are not the whole market. Seniority is inferred from job title text alone, and about two thirds of titles carry no level signal. "Associate" is deliberately excluded from the entry-level category because it is frequently used for mid and senior roles.

Licensed CC BY 4.0. Full method and the accompanying analysis are published at https://lastroundai.com/blog/salary-transparency-study-2026