Introduction

You publish a job to a hosted careers page or embed it on your own site, and every application arrives with its résumé parsed into the columns you screen on. A model reads each one against the job's requirements and answers with coverage and findings, every claim carrying the quote it came from, so a verdict can be checked instead of trusted.

Deciding stays yours — moving a candidate, writing to them, rejecting them — and each decision lands in a ledger you can read back. The same actions run from the krinoa CLI as JSON, which is what lets an agent drive the pipeline.

These docs cover the shape of Krinoa and the parts that take some setting up.

• Quickstart — post a job and read your first application.
• Concepts — the handful of words the product uses in its own way.
• Data and privacy — what is stored about a candidate, and what leaves.
• Setup — the screening model, WhatsApp, the careers page and the embed, referral invites.
• Agents & API — the CLI, pointing a coding agent at your workspace, the HTTP API.