A documented view of the observed tech hiring market.
Use the live aggregate snapshot, stable read-only API, CSV download, or Atom feeds. The public contract contains catalog summaries—not candidate data or full scraped vacancy descriptions.
Last reviewed Jul 29, 2026 · catalog-observation-v1
- Active roles
- 9,076
- Technologies
- 1,001
- Source dates
- 48.2%
151 observed companies.
Normalized catalog tags.
Of active roles include a source date.
Current aggregate snapshot
- 9,076 active and 708 archived vacancy observations.
- 151 companies and 1,001 normalized technology tags.
- 2.9% of active roles expose at least one salary boundary. Missing salary remains missing.
- Observation window currently spans Mar 8, 2013 to Jul 29, 2026.
Stable public API · v1
Start with /data/v1 for discovery or load the machine-readable OpenAPI 3.1 contract. Successful JSON responses use a versioned envelope with apiVersion, a five-minute generatedAt bucket, and data.
Read-only MCP server
Connect an MCP client to https://applydjinn.com/mcp. The stateless Streamable HTTP endpoint returns JSON responses and uses the same public-data projections as API v1.
Every tool is read-only and non-destructive. The MCP surface cannot access accounts, candidate profiles, recruiter personas, contacts, applications, documents, or full scraped descriptions.
- search_vacancies and get_vacancy for active public role summaries.
- technology_demand and compare_technologies for normalized demand evidence.
- company_hiring_activity for active role counts and leading technologies.
- market_snapshot for current coverage and data-quality indicators.
Data and privacy contract
Public vacancy records include structured summary fields, normalized technologies, an ApplyDjin canonical URL, and—on the single-record endpoint—the original source link. They do not include the full scraped job description.
No public data endpoint returns candidate profiles, recruiter personas, contacts, applications, application documents, account identifiers, email addresses, or private matching data.
Clients should send If-None-Match and respect cache headers. Version 1 is additive: fields can be added, while removal or incompatible meaning changes require a new major path.
How to interpret these numbers
The dataset describes the ApplyDjin observed catalog, not the entire labor market. Read the methodology and limitations before publishing comparisons or forecasts.
Research using these aggregates can cite ApplyDjin Research, the endpoint URL, methodology version, and retrieval date.