About Job Atlas Job Data

Four tools to start with

• LinkedIn jobs for professional-job alerts and trackers; YC jobs for startup communities; EURAXESS vacancies for research-career teams.
• AI Job Search & Fit Scorer for a developer-job shortlist with reasons, requirements, and gaps, for one candidate at a time.
• Versioned canonical, destination, and completion-summary contracts.
• Workflow and API starters for n8n, Make, MCP, REST, Zapier, Airtable, and Python. Templates contain no secrets; you add your own credentials. Zapier is a manual editor recipe.
• Open validation, responsible-use, quality-benchmark, and integration documentation.

Job Atlas and Nomad Agent

Job Atlas is the public product collection at jobatlas.dev. Its Apify listings use the job-atlas publisher. The existing domain, GitHub repository, and data-format identifiers retain their Nomad Agent names so existing links and integrations continue to work.

Choose a tool or find a workflow for your task.

What “open” means here

The schemas, scripts, Agent Skills, examples, integration assets, and documentation published in the repository are available under the MIT License. Hosted Actor implementations may be services and are not automatically included in that repository license.

The project does not claim that every scraper implementation is open source, every source is always available, or every workflow has been live-tested in every destination.

Independent integrations

Job Atlas is independently developed and is not affiliated with or endorsed by LinkedIn, EURAXESS, the European Commission, or source-site operators. Product and integration names identify compatibility and source context, not partnership.

Public page access is not authorization to crawl, retain, enrich, or republish data. Users are responsible for reviewing source terms, licensing, privacy obligations, and applicable law.

How decisions are made

Maintained integrations select latest and record the exact build used. They favor small first runs, conservative charge caps, stable identity, explicit unknowns, closed summaries, and named-destination checks. Marketing claims must be traceable to repository contracts, current public product metadata, or dated test evidence.

The public enrichment-quality benchmark does not yet report an official human-verified accuracy percentage. That number will be published only after the documented review process is complete.