Add freehire.me as an aggregator source — aggregated, but still the employer's own links

[strelov1]

Add freehire.me as an aggregator source

The README already offers a choice — scrape a career portal or use a job aggregator. freehire is the second kind, with the useful property that it crawls company career portals itself, so choosing it does not mean giving up the directness of the first option: every result keeps the employer's own ATS link.

freehire.me is an open-source IT job aggregator that crawls postings straight from company career boards and ATS platforms — 227 sources, 3.3M+ open roles across 294K companies — normalized into one schema and deduplicated. Live figures: /open.

GET /api/v1/jobs/search answers 200 signed out — no key, no cookie, no approval queue.

What a source adapter gets:

• Postings pre-tagged with stack, seniority, region, work mode, employment type and English level, plus a server-side visa/sponsorship filter
• Full descriptions as markdown inside the search response, so scoring needs no per-hit detail fetch
• The company's own ATS url on every result — the apply link stays the employer's, not an aggregator's
• Filter vocabulary discoverable at runtime via GET /api/v1/jobs/facets (canonical skill slugs, country codes, enums with live counts), so nothing is hardcoded against a vocabulary that changes

Links: OpenAPI https://freehire.me/openapi.yaml · docs https://freehire.me/docs/api · backend (Go, MIT) https://github.com/strelov1/freehire · CLI, --json on every command https://github.com/strelov1/freehire-cli

Scope note: IT/tech roles only, so it complements existing sources rather than replacing them.

Disclosure: I maintain freehire. Happy to write the adapter and open a PR if you want it — asking first rather than sending unsolicited code.

[vesaias]

Implemented and merged to main (4fa70f0) — thanks for the detailed writeup and the offer to help, it made this quick.

How it works: a new freehire search mode paginates GET /api/v1/jobs/search (public, no key) via limit/offset, and runs results through the same pipeline as every other source — title/company filters, body-exclusion/H-1B scan, salary extraction, and two-layer dedup. The employer's own url is kept as the apply link, and the markdown/HTML description is flattened to plaintext inline (no per-hit detail fetch, as you noted).

Configuration (Searches → mode "freehire.me"): paste a freehire search URL and its query filters (role, seniority, countries, collections, posted_within_days, …) are forwarded verbatim; an optional search term is sent as q and ANDed with the URL filters. The flask icon does a dry-run (filter diagnostics, no writes).

[vesaias]

Added salary in (c87447d) as well.