Indeed MCP Server

Give your AI agent live access to Indeed. This Actor is a Model Context Protocol (MCP) server that lets Claude, ChatGPT, Cursor, n8n, and any other MCP-compatible client search Indeed jobs, pull full job listings, and autocomplete locations and companies in real time — no scraping code required.

Prefer a classic scraper? The same Actor also runs as a one-shot dataset scraper when you give it a search query.

---

What it does

Expose Indeed to your AI agent as four tools:

Tool	What it returns
search_indeed_jobs	Job listings for a keyword + location: title, company, salary range, rating, location, snippet, apply link
get_indeed_job	Full detail for one job: complete description, precise salary, employment type, address, direct-apply flag
suggest_indeed_locations	Location autocomplete suggestions (e.g. "New York, NY")
suggest_indeed_companies	Company / keyword autocomplete suggestions

Works across 25 Indeed country sites (US, UK, Canada, Australia, India, Germany, France, and more).

---

Use cases

• AI job-search assistants — let an agent find and compare live openings for a user.
• Recruiting research — pull candidate-facing listings, salaries and company signals on demand.
• Salary benchmarking — ask your agent to gather salary ranges for a role across cities.
• Workflow automation (n8n / Make) — trigger Indeed lookups inside no-code automations.
• Classic scraping — run it without MCP to dump a job dataset for a query.

---

Connect it to your AI agent

1. Start this Actor in Standby mode (it runs as a persistent server).

2. Point your MCP client at the Standby URL with the /mcp path:

   https://<your-actor-standby-url>/mcp
   

3. Add your Apify API token as a header:

   Authorization: Bearer <APIFY_API_TOKEN>
   

Claude Desktop example

{
  "mcpServers": {
    "indeed": {
      "url": "https://<your-actor-standby-url>/mcp",
      "headers": { "Authorization": "Bearer <APIFY_API_TOKEN>" }
    }
  }
}

Then just ask: "Find remote senior Python jobs on Indeed and list their salaries and apply links."

---

Example tool call

search_indeed_jobs:

{
  "query": "software engineer",
  "location": "Remote",
  "country": "us",
  "limit": 20,
  "sort": "date"
}

Returns:

{
  "query": "software engineer",
  "country": "us",
  "count": 20,
  "results": [
    {
      "jobkey": "abc123",
      "title": "Senior Software Engineer",
      "company": "Acme Corp",
      "location_city": "Remote",
      "salary_min": 140000,
      "salary_max": 180000,
      "salary_currency": "USD",
      "salary_period": "YEARLY",
      "job_types": ["Full-time"],
      "company_rating": 4.1,
      "snippet": "We are hiring a senior engineer to ...",
      "url": "https://www.indeed.com/viewjob?jk=abc123"
    }
  ]
}

---

Classic scraper mode

Run the Actor normally (not Standby) with an input like:

{
  "query": "data analyst",
  "location": "London",
  "country": "uk",
  "limit": 100,
  "fetchDetails": true
}

It searches Indeed and pushes normalized job rows to the dataset.

---

Output fields

Each job includes: jobkey, title, company, company_rating, location_city, location_state, location_postal, salary_min, salary_max, salary_currency, salary_period, job_types, snippet, description, is_remote, apply_url, url, pub_date, create_date, scraped_at, and more.

---

Reliability & proxies

Indeed is protected by Cloudflare. For dependable results, run with Apify Residential proxies (the default). Anonymous access is best for the first result page per query — broaden coverage by varying the query, location and filters rather than deep pagination.

---

Notes

• This Actor returns publicly visible job-listing data. Respect Indeed's Terms of Service and applicable law; use the data responsibly.
• Not affiliated with or endorsed by Indeed.

FAQ

Do I need to write code? No. In MCP mode your AI agent calls the tools directly.

Which countries are supported? 25 Indeed country sites, selectable per call/run.

Can it get full descriptions and exact salaries? Yes — use get_indeed_job (MCP) or fetchDetails: true (classic mode).

Can I use it without MCP? Yes — run it as a normal scraper with a query.