Unofficial tool. Collect the open jobs of the companies you list from the public Greenhouse and Lever job board APIs: title, department, location, remote or hybrid, posting date, and salary where Lever publishes it. Schedule only-new runs for a daily feed of new openings. No recruiter data.
First version: reads the public Greenhouse Job Board API (/departments, or /jobs?content=true when descriptions are on) and the Lever Postings API (global and EU, 100 postings per page) for up to 500 boards per run.
One output shape for both APIs: company, title, department, team, locations, country, workplace type, employment type, posting and update dates, published salary range, job and apply links.
Filters: title keywords (any/all), department or team, location, workplace type, posted within N days.
"Only jobs not returned before" mode with a per-schedule state name, a run lock with heartbeat and a per-batch journal; returned jobs are remembered while listed and for 180 days after.
Privacy: no recruiter data; descriptions off by default; when on, plain text cut at 20,000 characters with e-mail addresses, phone numbers (North American and international), personal profile links and person-pointing sentences removed.
Polite access: one request at a time per API host with at least 1 second between requests, honest User-Agent, no proxies; a host that answers 401/403/407/429 is not called again in the run.
Pay-per-event with exactly three events: run-start (after the input passed validation), board (per board read with at least one job) and job (per saved job). The budget is checked before every charge, so a run never asks for more than the user's maximum cost.
Run state saved after every batch; migration and abort stop at a batch boundary, so no job or board is saved or charged twice.
