Gemini Live voice
Powered by Google Gemini Live API. Natural, low-latency conversation in 18+ languages. Configurable voice and model per bot.
One API call creates a session. Candidate speaks. You get scores, transcript, and a recommendation — delivered to your webhook.
# Create an assessment session curl -X POST https://leadvoice.botbrained.com/api/v1/sessions \ -H "Authorization: Bearer lv_sk_your_key" \ -d '{ "bot_id": "your-bot-uuid", "client_ref": "lead-001", "context": { "name": "Rahul Sharma", "course": "MBA" } }' # Response { "assessment_url": "https://leadvoice.botbrained.com/assess?t=…", "session_id": "uuid" }
LeadVoice plugs into any lead pipeline. No frontend code required on your end.
Write a system prompt in the admin UI. Set voice, language, and webhook. Use ${name} or ${course} for dynamic context.
POST to /api/v1/sessions with your API key, bot ID, and any lead context. Get back a secure, time-limited assessment URL.
Send the URL to the candidate — via WhatsApp, email, or SMS. They open it in any browser, pick their language, and speak.
When done, LeadVoice POSTs the transcript, scores (1–10), and a Hot/Warm/Cold recommendation to your endpoint — signed with HMAC-SHA256.
Everything you need to run voice assessments at scale.
Powered by Google Gemini Live API. Natural, low-latency conversation in 18+ languages. Configurable voice and model per bot.
Set the system prompt, persona name, voice, languages, and analysis model from the admin UI — no code deployment needed.
Assessment tokens are cryptographically random and never expose session IDs. Set expiry from 1 hour to 30 days.
Results delivered via HMAC-SHA256 signed POST. Automatic retry up to 5 times over an hour if your endpoint is down.
Generate multiple API keys per org, scoped to a specific bot. Revoke instantly from the admin UI. Keys are SHA-256 hashed — never stored in plaintext.
Sessions per day, completion rates, score distributions, and Hot/Warm/Cold breakdowns — all in the admin dashboard.
Pass lead data when creating a session. The bot's system prompt gets those values before the first word is spoken.
# In your system prompt: You are speaking with ${name}. They applied for the ${course} program in ${city}. Their budget is ${budget}. Open by greeting them warmly... # In your API call: { "bot_id": "...", "context": { "name": "Rahul Sharma", "course": "MBA", "city": "Mumbai", "budget": "50000" } }
Set up your first bot in under 10 minutes.
Get started →