Career Site Jobs MCP Server: Greenhouse, Ashby, Lever, Workday

Give your AI agent live job data straight from company career sites. This MCP server lets Claude, Cursor, ChatGPT or your own agent search open jobs, watch companies for new and closed roles, and summarize published salary ranges at any company that hires with Greenhouse, Ashby, Lever or Workday.

No login, no proxies, no browser. It reads the boards' public JSON APIs, so answers come back in seconds, and you pay only for what the agent uses.

Tools

Tool	What it does
search_jobs	Open jobs at up to 10 companies, newest first: title, location, remote/hybrid/onsite, department, salary range, posting date, apply URL. Filters by keywords, location, department, remote and age.
get_job_changes	Monitor a watch list. The first call saves a baseline; each later call with the same monitorId returns only jobs that are new, updated or closed since the previous call.
get_salary_stats	Yearly pay per currency (min, 25th percentile, median, 75th percentile, max) for matching jobs. Hourly and monthly ranges are converted to yearly.

Companies can be plain names (stripe, ramp), board URLs (https://jobs.lever.co/spotify, https://jobs.ashbyhq.com/linear, https://boards.greenhouse.io/airbnb) or source:name (lever:spotify). Names are looked up on Greenhouse, Ashby and Lever automatically. Workday always needs the career site URL, e.g. https://nvidia.wd5.myworkdayjobs.com/NVIDIAExternalCareerSite.

Connect

Through Apify's MCP server (easiest, works with any MCP client):

https://mcp.apify.com/?tools=nswk_api/career-site-jobs-mcp

Directly (Streamable HTTP), with your Apify API token as a Bearer token:

{
  "mcpServers": {
    "career-site-jobs": {
      "url": "https://nswk-api--career-site-jobs-mcp.apify.actor/mcp",
      "headers": { "Authorization": "Bearer <YOUR_APIFY_TOKEN>" }
    }
  }
}

Example prompts

• "Which of Stripe, Ramp and Linear are hiring senior backend engineers in Europe right now?"
• "Every morning, tell me which new ML roles opened at OpenAI, Anthropic and Mistral." (uses get_job_changes)
• "What do Datadog and Snowflake pay staff engineers in the US?" (uses get_salary_stats)
• "List remote product designer jobs posted in the last 7 days at these 10 startups."

Pricing

Pay per event, no monthly fee:

Event	Price
Company checked (one career site loaded)	$0.001
Job returned to the agent	$0.002
Job details lookup (Greenhouse salary or Workday description)	$0.001
Salary report (one get_salary_stats call)	$0.01

A typical search_jobs call on 3 companies returning 25 jobs costs about $0.05. Filtered-out jobs are never charged, and calls stop at your spending limit.

Output

Each job has jobId, title, company, source, location, countryCode, workplaceType, employmentType, departments, salary (min, max, currency, period), url, postedAt, updatedAt, changeType (in get_job_changes) and, on request, description (cut at 2,000 characters). Empty fields are left out to save the agent's context.

Limits and notes

• Up to 10 companies and 100 jobs per call; get_salary_stats analyzes up to 500 jobs.
• Large Workday sites (thousands of jobs, e.g. NVIDIA) take 1 to 3 minutes, because every job is listed before filtering. Raise your MCP client's tool timeout if needed.
• Salary ranges appear only where the company publishes them (common in the US, UK and Canada).
• Only public job postings are read. No personal data is collected.

Need a dataset or a scheduled feed instead of an agent tool? Use the single-source Actors: Greenhouse, Ashby, Lever and Workday.