Intervues · Product
A score nobody can explain is a score nobody should trust. Intervues attaches each rubric judgment to the specific moment in the transcript that produced it, so hiring managers and candidates can verify the claim.
Scores are applied against a structured rubric for the role — the same bar for every candidate in a cohort. That is what makes comparison fair: you are not comparing interviewer moods across Tuesday and Thursday panels.
Formal calibration against your own interviewers is part of every pilot. We do not publish a generic "85% agreement" number because your bar is yours.
Each score carries a citation: timestamp, quoted utterance, code snippet, or diagram element where relevant. Disputes get resolved against the transcript, not against a model's hidden state.
We do not claim a completed NYC Local Law 144 bias audit or SOC 2 certification. Underlying score distributions and calibration parameters exist for auditors — see bias audit for roadmap honesty.
Questions
Today: shareable report URLs, PDF export, and CSV import into custom ATS fields you configure. Native Greenhouse or Lever push is on the roadmap — see [ATS integrations](/product/ats-integrations).
Related: Pricing · Interview engine · Hiring teams
Run a structured screen on your JD and compare cited scores to your panel's judgment.
· 3 free credits · pay per interview · nothing recurring