— How it works
The pipeline is the product. Here's what runs between you finishing onboarding and you walking into an offer call.
We pull postings directly from the applicant tracking systems used by sponsoring employers. The refresh runs nightly, so what you see at 9 a.m. was posted overnight, not last week.
Source-of-truth ingestion means no scraped listings, no stale aggregator duplicates, and no expired jobs cluttering your shortlist.
We cross-reference each posting against 300+ companies on our internal sponsoring-employers list, derived from H-1B and LMIA disclosures plus a curated OPT/PGWP-friendly track record. Roles inherit a confidence label so you never wonder whether to apply.
Each posting is scored against your profile — skills, years of experience, role level, and location preference all feed the score. Anything below the threshold drops off your feed; anything above bubbles up with a one-line reason.
The Copilot runs four models: Claude Haiku 4.5 (fast, the default), GPT-5.4 (balanced reasoning), GPT-5.3 Codex (coding interviews), and Claude Sonnet 4.6 (the hard ones). Choose per interview from the Mac app.
The Copilot listens to the call, recognises the question being asked, and surfaces a 1–3 line outline based on the company context and your background. It's a confidence prompt, not a script. You answer in your own words. 10 hours a month, with top-ups when you need more.
— Max tier
Max is our delivery-team tier. Our team submits up to 15 applications a month on your behalf, you keep the Pro Copilot, and we run a monthly strategy call. $99/month, waitlist-only while we scale capacity.
Free to try. Pro adds the AI Interview Copilot — 10 hours a month, four models — for $39/month.