Company Hiring Signals: Greenhouse, Lever & Ashby Jobs API

Track competitor hiring, find companies that are scaling a team, and give your AI agent a "what is this company hiring for?" tool. Live open jobs from a company's own Greenhouse, Lever or Ashby job board, with department and location counts, new and closed roles since your last check, and a hiring-velocity score. Works as an MCP server for Claude, Cursor and other AI agents. From $0.01 per company.

What is a company hiring for right now, and is that changing? Give this Actor a company name, website or job-board link. It finds the company's own public job board on Greenhouse, Lever or Ashby and returns:

• every open role, normalised: title, department, team, location, remote flag, workplace type, employment type, posted and updated dates, a link to the original posting, and a short summary
• counts of open roles by department and by location
• new and closed roles since your last check (the Actor remembers the previous snapshot for each board)
• a simple, documented hiring-velocity score (0-100)

Use it for sales prospecting (teams that are scaling a function), investor and competitor research, recruiting market maps, or as a tool inside an AI agent (it is also an MCP server).

Why this Actor

• Official sources only. It reads the public job-board APIs that Greenhouse, Lever and Ashby provide for exactly this purpose. No logins, no scraping of LinkedIn or other job aggregators, no proxies.
• Change tracking built in. Run it on a schedule and each result tells you which roles appeared and which closed since the previous run.
• Cheap and predictable. Pay per company and per job returned. No start fee. Failed lookups are free. Set maxJobsPerCompany to 0 to pay only for the summary.
• Privacy aware. Recruiter and hiring-manager names, emails and phone numbers are removed from summaries, and descriptions are cut to a short summary with a link back to the source.

Input

Field	Description
companies	List of companies. Each entry can be a website (stripe.com), a name (Airbnb), an ats:token pair (lever:palantir, ashby:openai, greenhouse:airbnb) or a board URL (https://jobs.ashbyhq.com/openai). Up to 1000 per run.
companyInputs	Optional advanced form: objects with companyName, domain, boardToken, atsHint.
atsHint	Optional greenhouse, lever or ashby to check first.
includeJobs	Return the individual roles (default true). If false, only counts, changes and score are returned.
maxJobsPerCompany	Cap on roles returned per company (default 100). Counts and changes always cover the whole board.
summaryChars	Length of each role summary (default 280, 0 to omit).
keyword, department, location, remoteOnly	Filters for the returned role list. Counts are always for the whole board.
stateStoreName	Named key-value store used to remember previous snapshots (default hiring-signals-state). Use a different name to keep separate histories.

Example:

{
  "companies": ["stripe.com", "lever:palantir", "ashby:openai"],
  "maxJobsPerCompany": 25,
  "remoteOnly": false
}

Output

One dataset item per company. Abbreviated, illustrative example (numbers are not real):

{
  "input": "ashby:openai",
  "status": "ok",
  "company": { "name": "OpenAI", "domain": null },
  "board": {
    "ats": "ashby", "token": "openai", "boardUrl": "https://jobs.ashbyhq.com/openai",
    "method": "board_token", "confidence": "high", "alternatives": [], "notes": []
  },
  "fetchedAt": "2026-10-26T15:20:03.579Z",
  "openRoles": 790,
  "remoteRoles": 41,
  "countsByDepartment": { "Applied AI": 180, "Research": 95, "...": 0 },
  "countsByLocation": { "San Francisco": 520, "London": 40, "...": 0 },
  "changes": {
    "isFirstSnapshot": false, "previousSnapshotAt": "2026-10-19T15:20:00.000Z",
    "newCount": 22, "closedCount": 9, "netChange": 13,
    "newRoles": [{ "id": "...", "title": "...", "department": "...", "location": "..." }],
    "closedRoles": [{ "id": "...", "title": "...", "department": "...", "location": "..." }]
  },
  "velocity": { "score": 64, "label": "high", "components": { "freshness": 0.62, "scale": 1, "momentum": 0.5 } },
  "jobsReturned": 25,
  "jobsTruncated": true,
  "jobs": [{
    "id": "...", "title": "Research Engineer", "department": "Research", "team": null,
    "location": "San Francisco", "allLocations": ["San Francisco"], "remote": false, "workplaceType": "onsite",
    "employmentType": "FullTime", "postedAt": "2026-10-01T00:00:00.000Z", "updatedAt": null,
    "sourceUrl": "https://jobs.ashbyhq.com/openai/...", "summary": "Short summary, contact details removed."
  }],
  "attribution": "Data from the public ashby job board of openai. Each job links to its source posting; ..."
}

Companies that cannot be found are returned with status: "error" and a reason, and are not charged. The dataset has two views: Overview (one row per company) and Jobs (one row per role).

board.confidence tells you how the board was found: high when you gave a board token or the company's own careers page links to it, medium or low when it was matched by name. If a name exists on more than one ATS, the larger active board is chosen and the others are listed in board.alternatives.

Hiring-velocity score

A transparent heuristic, not a prediction. Version 1:

• freshness = share of open roles first posted in the last 30 days, scaled so 50% or more = 1.0 (neutral 0.5 if the board has no dates)
• scale = log10(1 + open roles) / log10(201), capped at 1.0 (200+ open roles = 1.0)
• momentum = net roles added minus closed over the last 30 days of stored snapshots, divided by the open-role count at the start of that window, mapped to clamp(0.5 + change, 0, 1). Neutral 0.5 until there is history.
• score = round(100 x (0.5 x freshness + 0.25 x scale + 0.25 x momentum))
• label: under 25 low, 25-49 moderate, 50-74 high, 75+ very high

Run the Actor on a schedule (for example weekly) so momentum reflects real change.

Pricing

Pay per event, no start fee:

Event	Price
Company snapshot (one company looked up successfully)	$0.01
Job returned (one role in the output)	$0.002

Examples:

• 100 companies, counts and changes only (includeJobs: false): $1.00
• One company with 50 roles: $0.01 + 50 x $0.002 = $0.11
• Weekly check of 25 competitors, summaries only: $0.25 a week, about $1.08 a month
• An agent asking about one company with the top 10 roles: $0.03

Apify's free plan includes monthly platform credit, which covers hundreds of company lookups. If you set a maximum cost per run, the Actor stops cleanly and trims the role list to fit.

Use with AI agents (MCP)

The Actor is an MCP server with one tool, get_hiring_signals (inputs: companyName, domain, boardToken, atsHint, includeJobs, maxJobs, keyword, department, location, remoteOnly). An agent can ask "What is Stripe hiring for in engineering?" or "Did Linear open any sales roles this week?" and get structured JSON back. There are two ways to connect.

Option 1: through the Apify MCP server (OAuth sign-in, easiest). Add this to Claude Desktop (claude_desktop_config.json), Cursor (.cursor/mcp.json), VS Code or any client that supports remote MCP servers:

{
  "mcpServers": {
    "hiring-signals": {
      "url": "https://mcp.apify.com/?tools=reynolds/company-hiring-signals"
    }
  }
}

Your client opens an Apify sign-in the first time. To use a token instead, add "headers": { "Authorization": "Bearer <APIFY_TOKEN>" }. Combine it with other Actors by listing them, for example ?tools=reynolds/company-hiring-signals,reynolds/domain-intel. Claude Code: claude mcp add --transport http hiring-signals "https://mcp.apify.com/?tools=reynolds/company-hiring-signals".

Option 2: connect straight to the Actor's Standby MCP endpoint (lowest latency, needs an Apify API token):

{
  "mcpServers": {
    "hiring-signals": {
      "url": "https://reynolds--company-hiring-signals.apify.actor/mcp",
      "headers": { "Authorization": "Bearer <APIFY_TOKEN>" }
    }
  }
}

A plain HTTP endpoint is also available: GET https://reynolds--company-hiring-signals.apify.actor/v1/hiring-signals?domain=stripe.com with the same header.

Clients that only support local (stdio) servers can use npx -y @apify/actors-mcp-server --tools reynolds/company-hiring-signals with APIFY_TOKEN set in the environment.

Integrations: API, n8n, Make, Zapier, LangChain

• REST API (run and get results in one call):

  curl -X POST "https://api.apify.com/v2/acts/reynolds~company-hiring-signals/run-sync-get-dataset-items" \
    -H "Authorization: Bearer $APIFY_TOKEN" -H "Content-Type: application/json" \
    -d '{"companies": ["stripe.com", "ashby:openai"], "includeJobs": false}'
  

• n8n: add the Apify node, choose "Run an Actor and get dataset", Actor reynolds/company-hiring-signals, and pass the input JSON. Put a Schedule trigger in front and a Slack or Sheets node after it to get a weekly "who started hiring" digest.

• Make and Zapier: use the Apify app's "Run an Actor" action with the same Actor and input, then read the dataset items in the next step.

• LangChain (Python):

  from langchain_apify import ApifyActorsTool
  hiring_tool = ApifyActorsTool("reynolds/company-hiring-signals")
  # give hiring_tool to your agent; it takes the Actor input as JSON
  

• Schedules: in Apify Console, create a Schedule that runs this Actor weekly with your company list. Each run reports roles added and closed since the previous one.

FAQ

Which companies are covered? Any company whose public job board is hosted on Greenhouse, Lever or Ashby. That includes a large share of venture-backed tech companies (for example Stripe, Airbnb, OpenAI, Palantir, Notion). Workday, SmartRecruiters and others are not covered yet.

Is this a LinkedIn or Indeed scraper? No. It reads the official public job-board APIs that the ATS vendors publish, so data comes straight from the company's own board and is current.

How do I track changes over time? Run the Actor on a schedule with the same stateStoreName. The first run stores a snapshot; every later run returns changes.newRoles and changes.closedRoles.

What does a lookup that fails cost? Nothing. Companies with status: "error" are not charged.

How do I keep costs low? Set includeJobs: false (or maxJobsPerCompany: 0) to pay only $0.01 per company for counts, changes and the score.

Can my AI agent call it? Yes. See "Use with AI agents (MCP)" above for Claude, Cursor and other MCP clients.

Data, privacy and terms

• Data comes only from the documented public job-board APIs of Greenhouse (boards-api.greenhouse.io), Lever (api.lever.co) and Ashby (api.ashbyhq.com). These are published by the ATS vendors so that companies' jobs can be displayed elsewhere.
• When a website is given, the Actor reads the company's own /careers, /jobs or home page only if robots.txt allows it, to find the board link.
• Requests are rate limited per host, identify themselves with an honest User-Agent, and use no proxies.
• Personal data: names of recruiters and hiring managers, email addresses and phone numbers are removed from summaries. Removal is automatic and best effort. Full descriptions are not stored; each role links to its source.
• You are responsible for using the results in line with applicable law and the job boards' terms.

Limitations

• Only Greenhouse, Lever and Ashby. Companies on Workday, SmartRecruiters, iCIMS, Teamtailor, Personio or a self-built system are not covered yet.
• Greenhouse boards hosted in the EU data centre are not covered yet.
• Matching by name can pick the wrong company when a short name is shared. Give a website or a board token for certainty, and check board.confidence.
• Change tracking starts from your first run; there is no back-filled history.

Local development

npm install
npm run build
npm test                 # unit tests (offline)
npm run test:live        # live checks against real boards
npm run smoke            # live smoke run, writes smoke-results.json / .md
npm run dev:cli -- --domain stripe.com --max 5
npm run start:mcp        # MCP over stdio
PORT=8787 npm run start:http   # HTTP API; HS_ENABLE_MCP_HTTP=1 also serves /mcp; X402_ENABLED=1 turns on the x402 stub

Environment: HS_CONTACT (contact added to the User-Agent), HS_MIN_INTERVAL_MS (default 250), HS_MAX_PER_HOST (default 2), HS_STATE_DIR (CLI and HTTP state, default .state).