An agent that speaks for your business has to be dependable. Here is how we protect it, your visitors and your leads.
TLS on every connection. Voice runs over encrypted WebRTC with a secure WebSocket fallback.
Each agent only accepts sessions from its approved domains. Copies on other sites will not connect.
Per-network rate limits, daily call caps, honeypot fields and input length limits on every endpoint.
Agents disclose they are AI, answer only from approved content and escalate instead of guessing.
Automated checks every 5 minutes with email alerts on failure and recovery. Live status →
No ad trackers. IPs hashed one way. Transcripts expire after 90 days, previews after 30.
The website is static and served from Vercel's global edge. Our API runs on Cloudflare Workers with a Cloudflare D1 database. Voice and conversation AI runs on ElevenLabs. No secrets ship to the browser and server tools never carry credentials in URLs.
Production access is limited to named team members. Business dashboards use long random keys that are stored in the browser, never sent in URLs to our servers and checked on every request.
We alert affected customers without undue delay and report to authorities as Indian law requires.
Email hello@livesupport.co with the subject "Security report". Please give us reasonable time to fix issues before disclosure. We will not pursue good-faith research. See also security.txt.