Tankha — Salary Data from Greenhouse, Lever, Ashby & More

What is Tankha?

Tankha ("salary" in Hindi) pulls salary data from public Greenhouse, Lever, Ashby and SmartRecruiters job boards. Give it a list of company boards. You get one row per job with the pay range, currency, pay period and the exact text the number came from. Run it again later and it tells you which jobs are new, which changed pay and which were removed.

Why salary data?

Pay-transparency laws in the US, Canada, the EU and the UK mean thousands of companies now publish pay ranges. Recruiters, compensation analysts, job seekers and market researchers use this data to benchmark offers, follow hiring trends and spot pay changes. Tankha reads the companies' own public job APIs, so the data is first-hand.

Supported boards

ATS	Write it as	Or paste the URL
Greenhouse	greenhouse:airbnb	boards.greenhouse.io/airbnb, job-boards.greenhouse.io/figma
Lever	lever:palantir	jobs.lever.co/spotify
Ashby	ashby:openai	jobs.ashbyhq.com/ramp
SmartRecruiters	smartrecruiters:BoschGroup	jobs.smartrecruiters.com/BoschGroup

How do I find a board name?

Open the company's careers page and click any job. If the address is on one of the domains above, the board name is the first part of the path, for example jobs.lever.co/spotify/…. Tankha reports an unknown or misspelled board in the summary and still runs the other boards.

How it extracts salaries

1. Structured fields first: Greenhouse pay_input_ranges, Ashby compensation tiers (including "Offers Equity") and Lever salaryRange.
2. Text parser second: a deterministic parser reads the job description ("Salary: $199,222 - $235,200/yr", "€94,000-€105,000 per year", "$32 - $35/hour").
3. Never guess: a range is reported only when there is a currency, a pay context, a clear pay period and plausible numbers. Tankha rejects revenue figures ("$250 - $1B revenue"), years, percentages, 401(k) and bonus amounts, funding and headcounts. Open ranges such as "up to $120k" keep the missing bound as null. A job with several regional ranges lists all of them in salaryRanges, and the first one is the primary.

India: LPA, lakh and crore

12-18 LPA, ₹12,00,000 – ₹18,00,000, INR 15 lakh, 1.2 crore and 12L-18L all become yearly INR numbers (for example 1200000-1800000).

Input example

{
  "boards": ["greenhouse:airbnb", "ashby:ramp", "lever:palantir"],
  "maxJobsPerBoard": 200,
  "onlyWithSalary": false,
  "trackChanges": true
}

Output example (real row, ashby:ramp)

{
  "source": "ashby", "board": "ramp", "company": "ramp",
  "title": "Security Engineer, Cloud",
  "url": "https://jobs.ashbyhq.com/ramp/34413f8d-26bf-4bbc-8ade-eb309a0e2245",
  "location": "New York, NY (HQ)", "department": "Engineering", "remote": true,
  "postedAt": "2026-04-07T17:12:35.753+00:00",
  "salaryDisclosed": true, "salaryStatus": "disclosed",
  "salary": { "min": 211400, "max": 290600, "currency": "USD", "period": "year",
    "source": "structured", "evidence": "$211.4K – $290.6K • Offers Equity", "equity": true },
  "changeType": "new", "previousSalary": null
}

The SUMMARY record in the key-value store has, for each board: job count, jobs with salary, disclosure %, the median, minimum and maximum annual salary per currency, change counts and any errors. In our test runs, Ramp disclosed pay for 96% of its jobs, Airbnb 84%, OpenAI 79% and Palantir 77%.

Change tracking

With trackChanges on (the default), Tankha saves a snapshot of each board in a named key-value store, tankha-snapshots, in your account. On the next run each job is marked:

• new: not seen last time
• salary_changed: the pay range differs; previousSalary holds the old range
• unchanged
• removed: the job is gone. These rows are free, and they are only reported when Tankha fetched the board's complete job list, so a failed or partial fetch never produces false removals.

Schedule the Actor daily or weekly to monitor competitors' hiring and pay.

Pricing

You pay per job analysed (the job-analyzed event). That is one event per job row delivered; removed rows are free. See the Pricing tab for the current price. To cap cost, use maxJobsPerBoard, or onlyWithSalary to receive and pay for only the jobs with pay data.

Limitations

• Tankha reads only the ATS's public API. Some companies show pay only on their own careers site. Stripe, for example, renders many US ranges on stripe.com, and those are absent from its Greenhouse feed. Such jobs come back as not_disclosed.
• A bare $ is read as USD unless the text names another dollar currency (CAD, AUD, SGD and so on).
• The text parser understands English plus common German pay phrases. Pay written only in other languages may be missed, never invented.
• On SmartRecruiters, each job needs one extra request for its description, so maxJobsPerBoard also limits requests there.

FAQ

Is this legal? Tankha reads the same public job APIs that power the companies' career pages. It does not log in or collect personal data.

Why is equity sometimes null? equity is true when the posting mentions equity or stock as part of pay. It is never set to false on a guess.

How are annual figures computed? Hourly × 2080, daily × 260, monthly × 12. They appear only in SUMMARY. Rows keep the original period.

Can I track hundreds of boards? Yes. Add up to 200 boards per run. Each board has its own snapshot.