Find an answer fast.

Sign in, click 'New job' on the recruiting dashboard, type a one-line description (e.g. 'Senior React engineer in Vienna, hybrid'), and PAIR drafts the JD. Edit or accept, then publish to your career page.

Six steps: branding, inbound email forwarding, your first JD, screening bot config, status-email templates, and adding your team. Most teams finish in 15 minutes. The 'Take the guided tour' button on the dashboard walks you through it.

Admin → Team & roles → Invite by email. Roles: Owner, HR Admin, HR User (no team management), Manager (own team only), Employee (self-service only).

JDs: paste the markdown into the JD editor (full bulk CSV import is on the roadmap). Candidates: forward your existing jobs@ inbox to your auto-provisioned PAIR inbound address — CVs are parsed and matched automatically. Set this up under Admin → Integrations → Inbound email.

Candidate applies → consent email goes out → candidate clicks the link and picks a slot → at that slot, our LiveKit-hosted AI (Anthropic Sonnet + Deepgram + ElevenLabs) calls them via Telnyx. Bot self-discloses as AI, asks the configured questions, captures recording + transcript, scores against the rubric, sends results to the application detail page.

Starter is chat-only. The screening still runs, just in the browser instead of by phone. Candidates type answers in a 10-minute conversation and the same AI scoring fires. Upgrade to Growth (€249/mo) for voice.

Yes. On every JD edit page, scroll to 'AI screening bot' → 'AI-draft from JD' generates 5–8 questions and a scoring rubric tuned to that role. Edit any question, weight, or knockout flag inline; save to apply.

Voice falls back to chat automatically. The screening still runs end-to-end; no manual intervention needed.

Yes. A nightly cron deletes recording files from disk; the metadata row stays for audit purposes but the audio is gone. Candidates can request immediate deletion any time via the GDPR portal at /c/gdpr.

The voice flow gracefully no-ops. JDs configured as voice/both fall back to chat screening. The dashboard shows an amber notice. Once Telnyx clears, voice flips on without any code change — just env vars in Coolify.

Yes. €49 / €249 / €599 covers your whole team — no extra charge per HR seat (within tier limits) or per employee. Voice minutes are the only metered thing, and overage rates are public on /pricing.

We don't cut you off. Overage bills at €0.30/min (Growth) or €0.20/min (Scale) on the next invoice. Hit 80% of your inclusion and we email a heads-up so it's never a surprise.

Yes. Upgrade is immediate (proration on this month's invoice). Downgrade takes effect at next billing cycle so you don't lose access mid-month.

30-day full refund on any paid subscription. After that, cancel any time but no pro-rata for the current month. No tricks, no fine print.

Yes. EU-hosted (Frankfurt), audit-logged for every AI decision, candidate consent captured before voice screening, recordings auto-deleted after 6 months, full data export and right-to-erasure via /c/gdpr. DPA template available; signed DPA standard on Scale and Enterprise.

Recruitment AI is classified as 'high-risk'. PAIR's stance: humans always make the final hire/no-hire decision (the AI score is a recommendation, never a gate); candidates can demand explanation of any AI scoring; emotion recognition in voice calls is banned by us, not just by law; bot identifies itself as AI on call pickup.

Hetzner Frankfurt for the app and Postgres database. ElevenLabs (US, covered by SCC) for TTS. Anthropic (US, covered by SCC) for LLM. Deepgram (US, EU region) for STT. Telnyx (EU) for the SIP trunk. All listed in the public DPA appendix.

By default: nobody outside your tenant — Postgres row-level security enforces tenant isolation. Pradeep (the founder) has database access for support and incident response, and that access is audit-logged. Hire automation never reads beyond your tenant scope.

Check spam, then your domain's DKIM/SPF setup. If you set up inbound email forwarding, make sure your forwarding rule keeps the original sender. If still stuck, reply to support@pair.careers with the candidate email address — we'll trace it in Resend.

Check the application detail page — the error message explains why. Common causes: candidate phone not E.164 (try +43… instead of 0660…), Telnyx number not yet active for outbound, or candidate's carrier rejecting the call. Click 'Schedule again' to retry.

Review the transcript on the application detail page — you can see exactly what the candidate said and how each criterion was scored. Per the EU AI Act, you can override any AI recommendation manually. Long-term fix: tune the rubric in the JD's screening bot config; weights are 0–1, knockout flags exist.

Candidate: application detail → GDPR tab → 'Delete all data' triggers full erasure across applications, screening sessions, voice calls, recordings, audit log entries. Employee: employee detail → settings → 'Delete'. Both are immediate and irreversible.