Click Open voice app and sign in from the panel. At the top of the voice page you’ll see Bot prompts (pick slot 1–12, edit the script, save). Use the Voice tab for ElevenLabs voice ID and clone upload. If the page looks old, hard-refresh (Ctrl+F5) or try a private window — CDN/browsers cache HTML.

Infrastructure → Tactic / Script Editor is separate: panel tactics for CDR / labels — not the AI Vishing App LLM prompts.

Faster TTS: on the voicecloner host, set PHONE_BOT_TTS_MODEL=eleven_turbo_v2 or eleven_turbo_v2_5 in .env (see .env.example), then docker compose up -d --build app. Shorter YAML prompts also reduce delay.

Choose devices here, then Open voice app — the link carries your selection (many headsets expose separate playback and capture devices). Grant mic access once if the browser hides device names.

Choose a project first. The Assignee list is loaded from Jira’s assignable users for that project. For call history, use CDR & Recordings and the Assignee filter there.

—