Y Combinator Jobs Scraper for Pipelines and Alerts

Get Y Combinator startup jobs for recruiting pipelines, job boards, and recurring alerts. Export complete descriptions, salary ranges, company details, and source links in a consistent format, with optional AI enrichment and cross-run deduplication.

Use this normalized Actor when you need repeatable alerts, structured job data, or records that share a format with the normalized LinkedIn and EURAXESS Actors. The separate legacy ycombinator-was-scraper serves existing flat-export workflows. Choose this Actor for nested records and pipeline controls; its stable source identifier keeps job keys compatible with existing YC records.

This is an independent service, unaffiliated with Y Combinator or Work at a Startup. It reads a regularly refreshed inventory of complete public postings; a run does not fetch YC pages on demand. Results reflect the last successful observation, and stale or unavailable inventory causes the run to fail.

Try five jobs

Open the Input JSON editor, paste the following, and run. Enrichment and translation are off, and deduplication is off so repeated trials can return the same jobs. This example costs $0.0045 in result events if five jobs arrive.

{
  "schemaVersion": "nomad-agent-job-search-input-v1",
  "firstRunMode": false,
  "postedWithin": "any",
  "maxItems": 5,
  "dedupe": {
    "enabled": false,
    "key": ""
  },
  "aiEnrichment": {
    "enabled": false,
    "accuracy": "silver"
  },
  "translateToEnglish": false,
  "includeRaw": true,
  "analyticsEnabled": false
}

Add "keyword": "founding engineer" or "location": "San Francisco" to narrow the search. Leave workplace filters blank for an initial trial. Download JSON from the Dataset tab; keep the nested record for downstream processing.

What a result looks like

This shortened excerpt comes from a delivered Knowtex job observed on 2026-09-04. It is a historical example, not a promise that the vacancy remains open. Fields omitted here are present in the full six-root record.

{
  "schemaVersion": "nomad-agent-job-v1",
  "identity": {
    "source": "ycombinator_was",
    "externalId": "107962",
    "url": "https://www.workatastartup.com/jobs/107962"
  },
  "data": {
    "title": "Engineer Manager, Platform",
    "company": {"name": "Knowtex"},
    "compensation": {
      "raw": "$230K - $280K",
      "minimum": 230000,
      "maximum": 280000,
      "exact": null,
      "currency": null,
      "period": null
    }
  }
}

Full records contain schemaVersion, identity, data, custom, llm, and raw. Company, salary, requirements, application links, and availability observations live in data. YC cohort, equity labels, source skill tags, and company context live in custom.data. The public YC v2 schema describes that extension. Raw description text and HTML remain available unless includeRaw is false.

Price examples

Current event prices, checked 2026-09-05:

Event	Per delivered job	Per 1,000 qualifying jobs
Job result	$0.0009	$0.90
Optional Silver enrichment	+$0.006	+$6.00
Optional Gold enrichment	+$0.010	+$10.00
Optional English translation	+$0.006	+$6.00

For 100 delivered jobs, base result events total $0.09. If all 100 also receive billable Silver enrichment, the total is $0.69; with translation on all 100, that becomes $1.29. Failed enrichment is not charged as enrichment. Translation is charged only for confirmed selected-field translation. These are Actor event examples; consult your Apify plan for any other applicable platform costs. Prices may change; the live Pricing tab is authoritative.

AI fills only missing fields supported by the description. It does not replace source facts. English translation is usually unnecessary for this largely English source and affects selected display fields, not the raw description. firstRunMode: true enables a paid five-result Silver-and-translation preset; leave it false for the base-price recipe above.

Build recurring alerts

Enable dedupe and use a stable non-empty key for one alert stream, for example {"enabled": true, "key": "my-yc-alert"}. Reuse that key and search scope on later scheduled runs to suppress previously delivered jobs. An empty key uses the Actor's default scope. Use identity.source + ":" + identity.externalId as the downstream job key; never deduplicate by job title or company.

postedWithin defaults to any. The supported 1h, 24h, 7d, and 30d windows use the hour when the inventory first saw the job, not a YC publication date. orderBy accepts newest or oldest using that same observation time. maxItems is capped at 1,000; zero requests that bounded window, not unlimited results. A bounded scan can return fewer matches and report resultsLimited.

For multiple terms, use ycSearch with schema version nomad-agent-ycombinator-search-v1, an OR-list in queries, and optional companyBatches, roleTypes, or jobTypes. Do not combine non-empty queries with keyword. See the input guide for a full recipe and normalized-filter reference.

Understand missing values

• null means unknown or unavailable; [] means explicitly empty.
• A dollar sign alone does not establish USD or an annual pay period.
• A location label can be available while parsed city and country remain null.
• A physical location does not establish onsite or hybrid work. Those filters can exclude most jobs with unknown workplace mode; use location when you mean a city label.
• First-observed dates and availability evidence describe past observations, not guaranteed publication dates or current hiring status.
• Founders are company context, not hiring contacts. Source skill tags are not automatically required skills.

API and AI agents

Use the YC guide and Agent Skill for a bounded API call using latest. Record the returned run’s immutable build ID and numeric build number. Read the run's RUN-SUMMARY (nomad-agent-run-summary-v4) and reconcile delivered with the dataset. resultsLimited describes incomplete coverage; a successful Apify status alone does not establish a complete search. Follow the summary's retry recommendation only within your authorized spend budget. The Actor never starts another paid run automatically.

Changelog

• September 2026: published the YC v2 extension schema and readable result links; clarified prices, alert setup, and source limitations. Preserved the accepted any recency default and existing normalized output semantics.

Job Atlas

Explore the Job Atlas job-data and matching Actors. Use latest and retain the immutable build ID returned by each run.

• Linkedin
• Euraxess
• YC
• Scorer

Website and integration guides | Source and client examples