Two hundred applications land overnight. The agent has read all of them, scored each one against your criteria, and put the eight worth your time at the top — before you open your laptop.
Most AI screening hands you a number and expects you to trust it. Jatura shows its working: the criteria it matched, the evidence it found, and what it couldn't verify.
Disagree with a row and re-score it. The agent takes the correction as a signal about your standards for that role, and the next batch reflects it.
How AI screening worksParses the resume, pulls out skills, experience and education, and matches them against the criteria you set on the job — not a generic keyword list.
Each candidate gets a score, a summary and the evidence behind it. You can disagree with the agent and re-score a row; it learns your standards from the corrections.
Sends the shortlist a self-scheduling link, checks it against your team's connected calendars, and handles the reminders and reschedules.
Status updates, follow-up questions and rejections go out over email or SMS from one inbox, so nobody sits in a black hole waiting on you.
Publish a keyword for a role. Candidates text it, the agent asks your screening questions over SMS, scores the answers, and drops them into the pipeline already ranked.
For frontline and high-volume roles this is the difference between a hundred half-finished web forms and a hundred screened candidates.
See text-to-applyScores are written to an audit trail with the model, the criteria and the reasoning captured at the time. Re-open any decision months later and see exactly what it was based on.
Anything the system decided without a human is recorded as such, alongside consent records and EEO data held apart from the hiring workflow.
The agent ranks and recommends. Advancing, rejecting and offering stay with your team — the DEI dashboard is there to show you where the pipeline is actually leaking.
AI work is metered in units — one unit is roughly 1,000 tokens. Starter includes 1,000 a month, Pro 5,000, Scale 25,000, and unused units roll over on Pro and above.
Run out and nothing breaks: the pipeline keeps running, you just don't get new AI scoring until the next cycle or a top-up. No surprise invoice, no throttling your hiring mid-week.
See plans and AI limitsJatura exposes an MCP server alongside the REST API and webhooks, so an assistant like Claude can query your pipeline, pull candidate context and take actions with a scoped API token — no scraping, no brittle integration layer.
API and MCP documentation