HN Who is Hiring — AI Structured Jobs

Every month, hundreds of companies post jobs in Hacker News' "Ask HN: Who is hiring?" thread. The posts are free-form text, so salary, remote rules, visa policy and tech stack are hard to filter.

This Actor uses AI (Claude) to turn each post into clean, filterable job rows — one row per role, even when a single post lists ten roles.

What you get

Field	Example
company_name, company_summary	Discord · "Communities and friend groups spend time together…"
title, seniority, category	Senior Software Engineer, Application Security · senior · security
work_mode, locations, remote_regions, timezone_note	onsite · San Francisco, CA
salary_min, salary_max, salary_currency, salary_period, salary_is_total_comp, has_equity	280000 · 330000 · USD · year · true
visa_sponsorship, citizenship_required, security_clearance	unspecified · null · null
employment_type	full_time / part_time / contract / internship
tech_stack	Python, TypeScript, React, Rust, Google Cloud Platform
apply_url, careers_url, company_url, contact_emails	links copied from the post
hn_url, posted_at, month	link back to the original comment
job_id	stable ID per role (<comment id>-<index>) for deduplication
first_seen_month, is_repost	whether the same company posted the same title in an earlier month

Rules the AI follows:

• No guessing. If the post doesn't say it, the field is null / unspecified. A plain "REMOTE" is not turned into "Worldwide".
• Links are verified. Every URL is checked against the original text; links that don't appear in the post are removed.
• Multi-role posts are split into one row per role, with shared salary/location copied only where the post applies it.

Why AI instead of regex

Regex parsers depend on the Company | Role | Location first line and return empty fields when a poster writes a paragraph instead. They also miss salary formats like 100-200k CHF or $280K–$330K TC. AI reads the whole post the way a person would.

Input

Option	Description
months	latest, all, or YYYY-MM (e.g. 2026-09). Several months allowed.
workModes	remote, hybrid, onsite, remote_or_onsite. For remote jobs use both remote and remote_or_onsite.
techStack	any of, exact name (case-insensitive), e.g. Python, PostgreSQL, Node.js. Add aliases when unsure (Rails, Ruby on Rails).
regions	any of, matched at word start in remote regions and office locations, e.g. EU, Berlin, Worldwide
keywords	any of, matched at word start in title, company, summary, tech stack (engineer matches Engineering)
requireVisaSponsorship	only posts that explicitly offer sponsorship
excludeCitizenshipRequired	drop roles requiring citizenship or clearance
seniority, categories, employmentTypes	any of the listed values, e.g. senior, data, full_time
minAnnualSalary, salaryCurrencies	e.g. 150000 and USD. Hourly/monthly pay is annualized. Roles without a salary are dropped.
postedAfter	2026-09-15 or relative like 7 days
excludeReposts	hide roles the company already posted in an earlier month
onlyNew	for scheduled runs: return only rows you haven't received yet with the same filters
maxItems	cap the number of rows (and your cost)

Example — remote Python roles open to the EU:

{
  "months": ["latest"],
  "workModes": ["remote"],
  "techStack": ["Python"],
  "regions": ["EU", "Europe", "Worldwide"]
}

Using it from an AI agent (MCP)

The Actor works well as a tool for AI agents via the Apify MCP server:

• A minimal call: {"months": ["latest"], "workModes": ["remote", "remote_or_onsite"], "techStack": ["Python"], "maxItems": 50}
• Every row has the same fields; missing information is null or "unspecified", never guessed.
• If nothing matches, the run's status message says so and lists the months that have data.
• Set maxItems (or a spending limit) to control cost — the run stops cleanly at your limit.

Coverage and freshness

• Months covered: every monthly thread from October 2025 onward (12+ months), and each new month is added automatically.
• Freshness: new threads are processed within a day of posting; late comments are added daily. Right after a new thread appears, latest keeps returning the previous month until the new one is ready.
• Tech names are unified: Postgres → PostgreSQL, NextJS → Next.js, Rails → Ruby on Rails, so techStack filters don't miss aliases.

Daily job alerts (scheduled runs)

Schedule this Actor daily with onlyNew: true and your filters to get only fresh roles each day, e.g.:

{
  "months": ["latest"],
  "workModes": ["remote", "remote_or_onsite"],
  "techStack": ["Python", "PostgreSQL"],
  "excludeReposts": true,
  "onlyNew": true
}

You pay only for new rows. Job IDs you already received are remembered in a key-value store named hn-who-is-hiring-state in your own account.

Pricing

$1.80 per 1,000 job rows. You pay only for rows that match your filters. Use maxItems to cap spend.

Runs are fast (a few seconds) because posts are parsed once in advance — your run only filters.

Output sample

{
  "month": "2026-09",
  "company_name": "Discord",
  "title": "Senior Software Engineer, Application Security",
  "seniority": "senior",
  "category": "security",
  "employment_type": "full_time",
  "work_mode": "onsite",
  "locations": ["San Francisco, CA"],
  "remote_regions": [],
  "salary_min": 280000,
  "salary_max": 330000,
  "salary_currency": "USD",
  "salary_period": "year",
  "salary_is_total_comp": true,
  "visa_sponsorship": "unspecified",
  "tech_stack": ["Python", "TypeScript", "React", "Rust", "JavaScript", "Google Cloud Platform"],
  "apply_url": "https://job-boards.greenhouse.io/discord/jobs/8656854002",
  "hn_url": "https://news.ycombinator.com/item?id=49525748"
}

Data source and responsible use

• Data comes from the official Hacker News API (public posts in the monthly "Who is hiring?" thread).
• contact_emails contains only addresses written out in plain text by the poster. Obfuscated addresses (e.g. name [at] domain [dot] com) are not decoded.
• Use contact details only to apply for the posted roles. Don't use them for bulk marketing. You are responsible for complying with applicable laws such as GDPR.
• AI extraction is very accurate but not perfect. Check the original post (hn_url) before acting on salary or visa details.

FAQ

Is "Who wants to be hired?" included? Not yet — this Actor covers the hiring thread only.

How are salaries normalized? 60-100k → 60000–100000. Currency is ISO 4217. TC sets salary_is_total_comp: true. Hourly rates use salary_period: "hour".

Found a wrong row? Open an issue with the hn_url and we'll fix the parser.

Related

• HN Who wants to be hired — AI Structured Candidates: the candidates' side — developers looking for work, with seniority, years, location, remote/relocation, visa, tech stack, email and résumé.