Y Combinator Companies Scraper: search the YC startup directory by batch, industry or tag

Pick a batch like W24, an industry, a region, a tag or a status, or paste company profile URLs. You get one row per company: the pitch, the long description, batch, status, team size, founded year, location, website, tags and the founders with their LinkedIn and X links.

The awkward bit is the founded year. It comes off the company's own profile page, and a fair number of profiles simply do not carry one, so yearFounded arrives null rather than estimated.

Input	Batches, industries, regions, tags, statuses, search terms or profile URLs
Output	One row per company
Ceiling	7,000 companies per run
Account needed	None
Price	$0.90 per 1,000 companies, flat on every plan

🏢 What Y Combinator Companies Scraper does

The YC directory is a searchable index of every company that has been through the programme, and each company also has its own profile page. This reads both and joins them into one row.

The directory card gives you the name, one-liner, batch, status, team size, industries, regions, tags, logo and website. The profile page adds the founders with their titles and bios, the year the company was founded, the city and country, and links to LinkedIn, X, Facebook, Crunchbase and GitHub. Every row has the same keys whichever way you searched.

Filters combine the way you would expect. Values inside one list are OR'd together, and different lists narrow each other. Ask for W24 and S23 with industries: ["B2B"] and you get B2B companies from either batch.

📥 What you give it

Every field is optional. Run it with the input empty and you get one labelled sample row back.

{
  "batches": ["W24", "S23"],
  "industries": ["B2B"],
  "statuses": ["Active"],
  "maxItems": 200
}

Field	Default	What it is
batches	none	Short (W24, S23, F24, P25) or long (Winter 2024). Up to 60. Each is queried separately, so ten batches and 100 rows gives ten from each.
industries	none	The directory's own labels: B2B, Consumer, Healthcare, Fintech and so on. Up to 60.
regions	none	The directory's region labels, like United States of America, Europe, India, Remote. Up to 60.
tags	none	The free-form profile tags, like Generative AI, Marketplace, Open Source. Up to 60.
statuses	all four	Active, Acquired, Public, Inactive.
searchTerms	none	Free text across names, pitches and descriptions. Each term is searched on its own.
companyUrls	none	Up to 500 profile URLs or bare slugs, to look companies up instead of searching.
allCompanies	off	Sweep the whole directory, batch by batch, newest first. Pair it with a high row limit.
isHiring	not applied	Three-state. See the note below.
topCompaniesOnly	off	Only companies the directory marks as top companies.
nonprofitOnly	off	Only the non-profits.
includeFounders	true	Reads each profile page too, which is what fills in the founders, founded year, city, country and social links. Turn it off for a lighter run. The price is the same either way.
maxItems	20	Total rows for the run, 1 to 7,000, shared across your batches and terms.
proxyUrls	none	Leave it empty for a normal run. It is there for callers who want traffic to leave through servers they already pay for, as http://user:pass@host:port.

isHiring has three states, not two. Leave it out and no hiring filter is applied. Set it true and you get companies flagged as hiring. Set it to false and you get only the companies that are not hiring, which is a legitimate query and almost never what somebody means. The Console writes a literal false into the input once you have ticked the box and unticked it again, so if a run comes back full of companies that are not hiring, remove the field rather than unticking it.

📤 What you get back

A real row, from run NctVtHlEJQFnIqU35:

{
  "ok": true, "charged": true, "recordType": "company", "query": "Winter 2024",
  "name": "Indemni", "slug": "indemni",
  "oneLiner": "Cargo Theft and Fraud Prevention Platform",
  "longDescription": "We are building a safer supply chain. Cargo Theft has been increasing yearly, ...",
  "batch": "Winter 2024", "batchCode": "W24", "status": "Active",
  "teamSize": 7, "yearFounded": 2024,
  "location": "San Francisco, CA, USA", "city": "San Francisco", "country": "US",
  "regions": ["United States of America", "America / Canada", "Remote", "Partly Remote"],
  "website": "http://www.indemni.com",
  "ycProfileUrl": "https://www.ycombinator.com/companies/indemni",
  "industry": "B2B", "industries": ["B2B", "Supply Chain and Logistics"],
  "subindustry": "B2B -> Supply Chain and Logistics",
  "tags": ["Identity", "Logistics", "Supply Chain", "Fraud Prevention", "Fraud Detection"],
  "founders": [
    { "name": "Omar Draz", "title": "Founder",
      "linkedinUrl": "https://linkedin.com/in/odraz",
      "twitterUrl": "https://twitter.com/oamdraz",
      "bio": "Ex-DoorDash, worked on Fraud, Growth and Logistics. Currently building!" }
  ],
  "founderNames": "Omar Draz", "founderCount": 1,
  "linkedinUrl": "https://www.linkedin.com/company/100487698/admin",
  "twitterUrl": null, "facebookUrl": null, "crunchbaseUrl": null, "githubUrl": null,
  "logoUrl": "https://bookface-images.s3.amazonaws.com/small_logos/a4f88ce8a0...png",
  "formerNames": ["Alacrity"],
  "isHiring": false, "nonprofit": false, "topCompany": false, "stage": "Early",
  "launchedAt": "2024-02-15T20:40:36.000Z", "companyId": "29533",
  "scrapedAt": "2026-09-21T01:52:29.035Z"
}

longDescription and logoUrl are cut short above. A real row carries both in full.

Field	What it is
companyId	The directory's own id, as a string. Use it as your dedupe key.
query	Which batch, term or filter produced this row, so you can tell where a result came from when you searched several things at once.
founders	Objects with name, title, linkedinUrl, twitterUrl and bio. founderNames is the same names flattened for a spreadsheet.
formerNames	Names the company has traded under before. Useful when your own list is out of date.
batch, batchCode	Winter 2024 and W24. Both are there so you do not have to convert.
stage	The directory's own label, like Early.
launchedAt	When the directory record was created, in ISO UTC. Not the founding date.

🧾 Reading the output

Three kinds of row.

Row	How to spot it	Billed
A company	recordType: "company"	yes
The sample row	_sample: true, and only when the input named no filter and no URL	no
A diagnostic	_diagnostic: true and ok: false	no

charged is set when the row is built, a moment before the charge goes out, so read it as "this is a real row" rather than as a receipt. A failed charge adds a CHARGE_ERROR diagnostic at the end.

The Console's default table hides the diagnostic columns, so those rows look blank there. Read them from the JSON.

Code	What it means
NO_RESULTS	The filter ran and matched nothing. An unrecognised batch string does this.
NOT_FOUND	A slug in companyUrls is not in the directory.
RATE_LIMITED, BLOCKED	The directory pushed back. Ask for less in one run.
NETWORK	A request could not be completed. Re-run it.
TIME_BUDGET	The run ran out of time before that slice was read.
PROXY_INPUT_ADJUSTED	Something in your proxyUrls was not usable and was adjusted.

▶️ How to run it

1. Open Y Combinator Companies Scraper and click Try for free.
2. Put batches in Batches, or terms in Search terms, or slugs in Company profile URLs.
3. Narrow it with Industries, Regions, Tags or Company status if you want.
4. Set Maximum companies, keeping it low on the first run.
5. Click Start, then download the dataset as JSON, CSV or Excel, or read it from the Apify API.

💰 How much does it cost?

$0.90 per 1,000 companies. The same rate on every Apify plan, with no volume tiers.

One charge per company row. The same company returned by two of your filters is only charged once. Sample rows and diagnostic rows are never charged. Reading each company's profile page for the founders does not change the rate: it is the same per company with includeFounders on or off.

💡 What people use it for

• Pulling a whole batch the week it is announced, for a newsletter or a tracker.
• Building a founder outreach list, since founders carries LinkedIn and X links directly.
• Filtering the directory for a thesis: a region, an industry and Active status together.
• Checking which YC companies in a space are hiring, then pairing it with a jobs scraper.
• Refreshing a stale list, using formerNames to catch companies that renamed.

🚧 What it does not do

• Companies only. No YC jobs, no Launch YC posts, no news, no funding rounds.
• No founder emails, and no contact details of any kind. Names and social links are what the directory publishes.
• yearFounded, city, country and the social links come from the profile page, so they are null when includeFounders is off.
• A profile page that fails to load still ships its row, and still bills, with founders empty and those profile-only fields null. Nothing on the row separates "could not read" from "not published", so re-run any slug that comes back suspiciously bare.
• One filter with no batches tops out at 1,000 matches, because that is as far as the directory's own search will page. Split the request by batch to get past it.
• 60 values per filter list, 500 URLs, 7,000 rows in a run. Anything past that is dropped.
• Team size and status are whatever the directory says today. They are not audited and they go stale when a company does not update its own record.

🧭 Which company scraper do you need?

If you want	Use
The Y Combinator startup directory	This one
New product launches and their makers	Product Hunt Scraper
Company pages on LinkedIn	LinkedIn Companies Scraper
Newly incorporated UK companies	Companies House New Companies Scraper
Who a startup is hiring	Wellfound Jobs Scraper

❓ Questions people ask

Do I need a YC login or an API key? No. Nothing to paste in and nothing to renew.

How do I get one whole batch? Put the batch code in batches and set maxItems above the batch size. A recent batch runs to a few hundred companies.

Can I get the entire directory? Tick allCompanies and raise maxItems. It walks batch by batch, newest first, and you pay per company, so decide the number before you start it.

Why did one filter stop at 1,000? Because the directory's search pages that far and no further. Add batches and the run splits the query per batch, which gets around it.

What is query for? It names the batch, term or filter each row came from, which is the only way to attribute results when you ran several searches in one go.

Is this legal? The directory and the profile pages are public. Rows carry founders' names and links, which data-protection law covers, so have a reason for collecting them. Apify's write-up on scraping and the law is a fair start, and we are not lawyers.

🆘 If something breaks

Open the Issues tab on the actor page. Send the run ID and the input you used. A diagnostic row's errorCode and the query field on it usually point straight at the filter that went wrong.