Wellfound Startup Salary Benchmarks by Role and City

Reads Wellfound's own startup hiring data. Salary percentiles by role, by city, by industry, by skill and by seniority, plus the quarterly series behind them. No key and no login.

The numbers are thousands and this Actor expands them

Wellfound stores a median of 125 and prints it as "$125k". Read the stored value and you get 125, which looks like a salary of a hundred and twenty-five dollars and sorts like one.

Every figure here comes out in full dollars. The columns are salaryP25Usd, salaryP50Usd and salaryP90Usd. Industry percentiles have halves in them, so 287.5 becomes 287500 and is not rounded to the nearest thousand.

Wellfound quotes every location in US dollars, London included.

The second line on the chart is not your role

Every role page carries two quarterly series. One is the role you asked for. The other is Wellfound's market-wide line, and its query carries your role while its numbers ignore it. The same figures come back on the overall page, on a role page and on a city page.

Those rows are marked isBaseline true and their scopeKind is all. Drop them unless you want the comparison line.

The same applies to the tables. On a city page the by-location table is the global one, not a breakdown of that city. Every row carries the scope its own table was computed for, read from the query rather than from the page it sat on, so scopeSlug and requestedScopeSlug can differ and both are on the row.

A role slug it does not know

Ask for /hiring-data/r/not-a-role and Wellfound answers with HTTP 200 and its overall figures, with no roleId anywhere in the query. Nothing looks wrong. You would be reading the whole market's pay as one role's pay.

This Actor checks the query for the role or city id, drops the page when it is missing, says so in the run summary, and charges you nothing for it.

How old the data is

Wellfound titles these pages with the current year. The quarterly series does not reach it. Read the run summary for earliestQuarter and latestQuarter, and every row carries periodStart and observedAt so you can see the gap yourself.

Wellfound publishes no sample size, so a thin market can produce a strange figure. A data scientist median of 205,000 in New York sits in the same table as 10,000 in Vancouver. Both are what the source says.

Equity

There is none here. The page title says salary and equity. The data behind it carries salary percentiles only, so this Actor returns salary.

Input

Field	What it does
roles	Role slugs from the Select role dropdown on wellfound.com/hiring-data
locations	City slugs, for example new-york, london, remote-friendly
includeOverall	Also read the overall page, every role and every city together
breakdowns	Keep only some tables. Empty means all of them
maxResults	Upper bound for the whole run

Output

One row per figure. A role page gives about 62 rows: 12 quarters for the role, 12 for the baseline, and the city, industry, skill and seniority tables.

RUN_SUMMARY in the key value store holds the pages read, how many figures each gave, the earliest and latest quarter, and whether any role or city was not recognised.

Price

0.002 dollars per run and 0.002 dollars per figure. Both are charged after the rows exist, so a run that finds nothing costs nothing.