Wellfound Jobs Scraper: Startup Jobs to CSV & JSON

Export public Wellfound startup jobs by role, location or job URL. Get one Dataset row per unique job, with stable IDs, salary text, posting date and description, ready for CSV, JSON or your job board.

• What it does
• How it works
• Quick start
• Use cases
• Input
• Output
• Pricing
• Limits
• API, export and integrations
• Troubleshooting
• FAQ
• Related Actors
• Support

What it does

The Actor reads public Wellfound job search pages and job pages and saves each startup job as one clean row: title, company, locations, remote flag, employment type, published pay, posting date, description and job URL. It reads only what Wellfound shows without a login. It is not a full index of every Wellfound job.

Good fit

• Track startup hiring for one role in one city, for example software engineers in New York.
• Feed a job board or newsletter with fresh startup jobs and their links.
• Compare published salary ranges for a role across startups.

👉 Try it now: open the input form

How it works

1. You give a Wellfound search URL, or a role and a location, or one job URL.
2. The Actor opens the public pages, removes duplicate jobs and applies your keyword, remote, job type and posting-age filters.
3. It saves one row per job to the Dataset and a short run report (RUN_SUMMARY) with the status and any problems.

👉 Try it now: open the input form

Quick start

1. Open the Input tab. The form already has a Wellfound search URL for software engineers in New York (box 1 in the picture).

2. Keep the small limits for your first run: Maximum items 3, Maximum pages per URL 1, Maximum request attempts 8 (box 2).

3. Click Start (box 3).

4. Open the Output tab. Check Job title and Company (box 1), Compensation, Posted and Job URL (box 2). Click Export (box 3) to download CSV, JSON or Excel.

   

Expected result: three software engineering jobs in New York with company, pay range, posting date and link, in about 5 seconds. Jobs on Wellfound change, so your rows will differ.

The form values are a prefill, not API defaults. An API input {} reads the public /jobs feed with maxItems: 50, three pages and descriptions on.

👉 Try it now: open the input form

Use cases

Buyer job	Input path	Output entity	Key fields	Next integration
Startup hiring in one city	roles + locations	job row	companyName, title, locations	CSV export to a spreadsheet
Job board or newsletter feed	startUrls, maxItems	job row	jobId, title, jobUrl, postedAt	Dataset API, scheduled task
Remote startup jobs	remoteOnly: true	job row	remote, remoteKind, acceptedRemoteLocations	CSV or JSON export
Salary research	small role search	job row	compensationText, salaryMin, salaryMax, salaryCurrency	spreadsheet pivot
Check one known job	one job URL, mode: exact_job	job row	description, postedAt, jobUrl	RUN_SUMMARY status

👉 Try it now: open the input form

Input

No field is required. Roles and locations are Wellfound URL slugs (lowercase letters, numbers and hyphens, for example software-engineer, new-york), not free text. Invalid limits, types or slugs stop the run with INVALID_INPUT before any page is read. In a mixed list, a bad URL is reported and the valid URLs still return rows.

Input fields

Field	Plain name	Type	Default / prefill	Allowed values	Effect
mode	Mode	string	auto	auto, search, exact_job, company_jobs	auto detects each URL; other modes require a matching URL.
startUrls	Start URLs	array of {url}	[]; prefill: New York search	HTTPS URLs on wellfound.com or legacy angel.co	Pages to read. Empty and no role/location means the public /jobs feed.
roles	Roles	string[]	none	up to 5 slugs	Adds /role/r/<role> pages, or every role × location pair.
locations	Locations	string[]	none	up to 5 slugs	Adds /location/<slug> pages, or every role × location pair.
remoteOnly	Remote only	boolean	false		Keeps only jobs the source marks remote; unknown is excluded.
query	Keyword	string	""		Keyword match in fetched title, company, role and description. Not a site-wide search.
jobType	Employment type	string	any	any, full-time, part-time, contract, internship, cofounder	Filters fetched jobs; unknown type does not match.
postedWithinDays	Posted within days	integer	none	0–3650	Keeps jobs posted in the last N days; unknown dates do not match.
includeDescription	Include descriptions	boolean	true		Opens one job page per row when the search page has no description.
maxItems	Maximum items	integer	50; prefill 3	1–1000	Rows for the whole run.
maxPagesPerSeed	Maximum pages per URL	integer	3; prefill 1	1–50	Search pages per URL.
maxRequests	Maximum request attempts	integer	200; prefill 8	1–1500	All page requests, including retries and descriptions.
requestTimeoutSecs	Request timeout seconds	integer	20	5–60	Time for one page.
requestDelayMs	Request delay milliseconds	integer	750	500–10000	Pause between requests, plus up to 25% jitter.
maxRetries	Transient retries	integer	2	0–2	Retries for timeouts, 429 and some 5xx. Never for 401/403/404.
transport	Transport	string	AUTO	AUTO, DIRECT	Both read pages directly over HTTPS.
deliveryConnector	Send results to (optional)	MCP connector	none	your Slack or Notion connector	Sends a short run summary after the data is saved.
deliveryTarget	Notion page or channel (optional)	string	none	Notion page link or Slack channel ID	Where the summary goes.

Interacting settings

When	Input A	Input B	Effective behavior
Both URLs and slugs	startUrls	roles / locations	Combined and deduplicated, at most 25 unique pages to start from.
Roles and locations	roles	locations	Every role × location pair becomes one search page.
Filters	query, remoteOnly, jobType, postedWithinDays	maxItems	Collection stops at maxItems unique jobs before filters, so a narrow filter can return fewer rows.

Input examples

Remote software jobs in New York

{
  "roles": ["software-engineer"],
  "locations": ["new-york"],
  "remoteOnly": true,
  "maxItems": 20,
  "maxPagesPerSeed": 1,
  "maxRequests": 8
}

Expected output: up to 20 job rows with remote: true.

Check one job

{
  "mode": "exact_job",
  "startUrls": [
    { "url": "https://wellfound.com/jobs/3628014-software-engineer" }
  ],
  "maxItems": 1,
  "maxRequests": 3
}

Copy a fresh jobUrl from a search run: a job can close after you copy it.

👉 Try it now: open the input form

Output

The Dataset has job rows only, no error or billing rows. Every field is always present: unknown text and numbers are null, unknown lists are []. Nothing is guessed: no salary, date or remote flag is filled in when the source does not show it.

Output fields

Field	Meaning	Type	Can be empty?
recordType	always JOB_LISTING	string	no
id	wellfound:<jobId>	string	no
jobId	Wellfound numeric job ID, as text	string	no
title	job title	string	no
slug	job slug from the URL	string	yes
jobUrl	https://wellfound.com/jobs/<id>-<slug>	URL	no
companyId, companyName, companySlug, companyUrl	public employer details	string / URL	yes
primaryRole	role label shown by the source	string	yes
jobType	full-time, part-time, contract, internship, cofounder	string	yes
locations	listed work locations	string[]	empty list if none
acceptedRemoteLocations	regions the job hires remotely from	string[]	empty list if none
remote	source marks the job remote	boolean	yes
remoteKind	REMOTE, ONSITE_OR_REMOTE, ONSITE	string	yes
wfhFlexible	source work-from-home flag	boolean	yes
compensationText	pay as published, for example $176k – $204k	string	yes
salaryMin, salaryMax	parsed pay range, only when clear	number	yes
salaryCurrency	USD, EUR, GBP or source code; never converted	string	yes
equityMinPercent, equityMaxPercent	equity range in percent	number	yes
compensationParseStatus	not_present, parsed, partial, ambiguous	string	no
experienceMinYears, experienceMaxYears	required experience	number	yes
postedAt	posting time, ISO	string	yes
postedAtSource	liveStartAt, jsonld, source_text, absent	string	no
description	public job description	string	yes
descriptionFormat	markdown, html_text, absent	string	no
atsSource	public applicant-tracking label; not opened	string	yes
badges	public company and job labels	string[]	empty list if none
sourceUrl, sourceMode, sourcePage	page that produced the row	mixed	sourcePage yes
observedAt	time the row was read, UTC	string	no
coverage	always public_page_subset	string	no

Pay keeps the source scale and period: it is not turned into a yearly amount or converted between currencies. Equity is in percent, not a fraction.

Output entities and surfaces

Situation	Where	What you see	What to do
Normal run	Dataset, RUN_SUMMARY	job rows; status: SUCCEEDED	export CSV or JSON
No matching jobs	RUN_SUMMARY	empty Dataset; status: EMPTY	loosen filters or pick another role/location
Some pages failed	Dataset, RUN_DIAGNOSTICS	rows kept; status: PARTIAL	read the diagnostic codes
Blocked or invalid	RUN_SUMMARY, RUN_DIAGNOSTICS	status: FAILED, run exits with an error	see Troubleshooting

RUN_PROGRESS shows the current page and counters while the run is going. A green platform status is not enough: check RUN_SUMMARY.status.

Output example

One row from the Quick start run on 3 November 2026, description shortened:

{
  "recordType": "JOB_LISTING",
  "id": "wellfound:3628014",
  "jobId": "3628014",
  "title": "Software Engineer",
  "jobUrl": "https://wellfound.com/jobs/3628014-software-engineer",
  "companyName": "January",
  "jobType": "full-time",
  "locations": ["New York"],
  "remote": false,
  "compensationText": "$176k – $204k",
  "salaryMin": 176000,
  "salaryMax": 204000,
  "salaryCurrency": "USD",
  "postedAt": "2026-10-19T16:30:33.000Z",
  "descriptionFormat": "markdown",
  "sourceMode": "search",
  "coverage": "public_page_subset"
}

Send results to Slack or Notion

Pick your connector in the optional deliveryConnector field. After the run saves its data, the Actor posts a short summary (up to 25 rows and a link to the full dataset) there. Put a Notion page link or a Slack channel ID in deliveryTarget: in Notion every run is added on top of that one page, so your workspace does not fill up with new pages. A delivery error never fails the run and costs nothing extra. Connect a service once in Apify Console → MCP connectors. The result is in RUN_SUMMARY.delivery.

👉 Try it now: open the input form

Pricing

You pay per result. Platform usage is included in the price.

Event	Free	Bronze	Silver	Gold+
Job listing saved to the Dataset, per 1,000	$0.25	$0.225	$0.20	$0.175
Actor start, per GB of memory	$0.0133	$0.0133	$0.0133	$0.0133

• The start fee is charged once per started GB of run memory, at least once. The default 256 MB run pays it once.
• Example: 100 jobs on the Free plan with default memory cost about $0.0133 + 100 × $0.00025 = $0.038.
• Empty, filtered, duplicate and failed results have no row fee. A run that saves nothing pays only the start fee.
• Set Maximum cost per run in Run options to cap what one run can spend.

👉 Try it now: open the input form

Limits

• Supported pages: /jobs, /role/r/<role>, /role/l/<role>/<location>, /location/<slug> and job pages /jobs/<id>-<slug>. Only wellfound.com and legacy angel.co HTTPS URLs; up to 5 redirects inside Wellfound.
• Company job pages (/company/<slug>/jobs) often return a login wall (HTTP 403). They are accepted and reported, but company coverage is not promised. Use a search or job URL instead.
• Up to 1,000 rows and 25 start pages per run. This is a cap, not a promise: pages stop when the source runs out, repeats jobs or hits your limits.
• One search URL gave about 49 unique jobs in our tests; later pages repeated the same jobs. For more jobs, add several role or location URLs.
• By our estimate one run makes about 200–250 page requests in 240 seconds at the default delay. A job whose search card has no description takes one more request, so several URLs together give up to about 200 jobs with descriptions per run (an estimate, not a tested maximum).
• Source work stops after 240 seconds even with a longer run timeout; rows found before that are kept with PARTIAL. The run needs more than 15 seconds left to start.
• maxRequests accepts up to 1,500 (default 200), but the 240-second limit usually stops a run first.
• Busy or failing pages (HTTP 429 and 5xx) are retried a few times with a pause.
• One page may be up to 8 MB; larger pages are skipped with BODY_TOO_LARGE.
• Memory 256–512 MB, default 256 MB. The Actor uses about 120 MB, so 512 MB does not make it faster; it only doubles the platform cost of the run.
• Results are a sample of public pages (coverage: public_page_subset), not every Wellfound job. Jobs, order and fields can change between runs.
• No login, cookies, proxy, browser or CAPTCHA solving.
• Each run is a fresh snapshot: there is no "only new jobs" mode. Compare jobId in your own storage.

This tool is independent of Wellfound and AngelList. Use the data in line with the Wellfound terms and the rules that apply to you.

👉 Try it now: open the input form

API, export and integrations

• API: start a run with the same JSON input as the form. A small run:

  curl --request POST \
    'https://api.apify.com/v2/acts/cliqtomedia~wellfound-jobs-scraper/runs?memory=256&timeout=300&maxTotalChargeUsd=0.05' \
    --header "Authorization: Bearer $APIFY_TOKEN" \
    --header 'Content-Type: application/json' \
    --data '{"roles":["software-engineer"],"locations":["new-york"],"maxItems":3,"maxPagesPerSeed":1,"maxRequests":8}'
  

  Then read the run's default Dataset. See the Run API.

• Export: the Dataset Export button gives CSV, JSON, Excel and other formats.

• Scheduling: save your input as a task and add a schedule in Apify Console. Use jobId to compare snapshots.

• Slack or Notion: see Send results to Slack or Notion.

Use with AI agents (MCP)

• MCP server with only this Actor: https://mcp.apify.com?tools=cliqtomedia/wellfound-jobs-scraper.
• Smallest useful input: {"startUrls":[{"url":"https://wellfound.com/role/l/software-engineer/new-york"}],"maxItems":3,"maxPagesPerSeed":1,"maxRequests":8}.
• Key output fields: title (job title), companyName (employer), compensationText (pay as published), postedAt (posting time), jobUrl (link to apply).
• The Quick start input takes about 5 seconds and costs about $0.014 on the Free plan. Keep maxItems small; one run returns at most 1,000 rows.

👉 Try it now: open the input form

Troubleshooting

RUN_SUMMARY.status tells how the run ended:

• SUCCEEDED: all pages were read.
• PARTIAL: rows are saved, but something was skipped, for example one job description could not be read (DETAIL_FETCH_FAILED) or the time or request limit was reached. The saved rows are valid.
• EMPTY: the pages worked but no job matched. Loosen filters.
• FAILED: nothing could be read. Check the codes below.

Code or symptom	What it means and what to do
SOURCE_BLOCKED or login wall	Wellfound refused the page. Use a public search or job URL. Do not add cookies.
SOURCE_EMPTY	The page had no jobs. Try a wider role or location.
HTTP_404	The job was closed or the ID is wrong. Copy a fresh jobUrl.
HTTP_ERROR	Wellfound returned an error after retries. Try again later.
TIMEOUT	A page or the 240-second source limit ran out. Lower maxItems or pages.
REQUEST_BUDGET_EXHAUSTED	maxRequests was used up. Raise it or narrow the search.
BODY_TOO_LARGE	A page was over 8 MB and was skipped.
REDIRECT_OUTSIDE_ALLOWLIST or REDIRECT_LIMIT	A link left Wellfound or redirected more than 5 times. Use a direct Wellfound URL.
MALFORMED_STATE or MISSING_STABLE_ID	The page layout changed or a job had no ID. Valid rows are kept; report the run ID.
DETAIL_FETCH_FAILED	One job description could not be read. The row is saved without it.
DUPLICATE or FILTERED	Repeated or filtered jobs were skipped. They are not charged.
INVALID_INPUT	Check slugs, limits, URL and mode. Start from the Quick start input.
INVALID_RESOURCES	Use 256–512 MB memory and leave more than 15 seconds of run time.
TRANSPORT_UNAVAILABLE or INTERNAL_ERROR	A temporary run problem. Start again; if it repeats, report the run ID.
CHARGE_LIMIT_REACHED	Your maximum cost per run was reached. Raise it or lower maxItems.
Fewer rows than maxItems	Filters, page limits and source end can stop early. A larger limit cannot remove a login wall.
Missing pay, remote or description	null means the source did not show it.

👉 Try it now: open the input form

FAQ

Do I need a Wellfound account or a proxy?

No. The Actor reads public pages without login and without a proxy.

Does the keyword search all Wellfound jobs?

No. query filters the jobs read from your search pages. Choose the role and location to set the search window.

Why do some jobs have no salary or date?

The source did not show them. Unknown values stay null; nothing is guessed.

Will a repeat run return only new jobs?

No. Each run is a new snapshot and each saved row is charged again. Compare jobId in your own storage.

Is this an official Wellfound API?

No. It is an independent tool that reads public pages.

👉 Try it now: open the input form

Related Actors

• YC Companies Scraper — Y Combinator startups with founders and open jobs.
• Remote Jobs Aggregator — remote jobs from RemoteOK, We Work Remotely and Himalayas.
• Google Jobs Scraper — jobs from many boards with salary and apply links.
• Ashby Jobs Scraper — many startups post jobs on Ashby job boards.
• Lever Jobs Scraper — jobs from public Lever boards.

👉 Try it now: open the input form

Support

For a bug, question or feature request, open the Issues tab of this Actor on Apify. Include the run ID, product status and diagnostic code. Never post tokens, cookies or private data.

👉 Try it now: open the input form