Job description in
Paste the JD. The agent extracts a structured requirement list — each one weighted, and flagged must-have or nice-to-have.
not on hunches
Paste a job description. Get candidates scored against it with a cited quote behind every verdict, an adaptive AI interview that probes only the gaps, and a feedback report drafted for you to approve.
How it works
The same quotes that justify a score are the quotes the interview probes, and the same quotes the rejection letter cites. Nothing is re-derived and nothing is invented.
Paste the JD. The agent extracts a structured requirement list — each one weighted, and flagged must-have or nice-to-have.
Every requirement gets a verdict on a 4-point scale, and every verdict is backed by a literal quote highlighted in the resume. Never an unexplained number.
Whatever scored Partial or Unproven becomes the input to an AI interview that asks about those gaps specifically — and is built to resist AI-copilot cheating.
If rejected, an auto-drafted report explains why, citing the same evidence — reviewed and approved by the recruiter before it reaches the candidate.
Why it’s different
Recruiters get proof instead of a ranking. Candidates get an answer instead of silence.
Every verdict cites a real quote from the resume. Click through and see it highlighted in place, in the original document.
Designed against AI-assisted cheating: consistent response-latency detection, on-screen artifact questions a copilot cannot see, and focus/tab monitoring.
Rejected candidates get a real, specific explanation instead of silence — drafted by AI, approved by a human before it sends.
Where it stands
These are the numbers the build actually produced. There are no customer counts here because there are not any yet.
Across the test resumes used during development, every citation pointed at text that actually exists in the resume.
Job description → evidence-backed scoring → adaptive interview → candidate feedback report.
Strong, Partial, Unproven, Absent. A recruiter can read a score and know exactly what it means.
Built for AI Agent Hackathon 2026 (Product Space) — September 2026.
Start with one job description and a stack of resumes. You will see the quotes before you see a shortlist.
Get started