Privacy Policy
Last updated: April 26, 2026
This policy describes how Natter ("we", "us") handles personal data when you sign in, create voice agents, and place or receive phone calls through our service. Read it carefully. Email hello@natter.ai with any questions.
What we collect
- Account info. Your email address (via Firebase Authentication) and a session cookie that identifies you across requests. We never store passwords — sign-in uses email-link verification.
- Voice agent configuration. The scripts, agents, and inbound routing rules you create. Stored in Google Firestore (project
gnbly0) under your user ID. - Call records. When you place or receive a call through Natter, we store call metadata (duration, cost, status, the phone numbers involved), transcripts, and (if you enable recording) audio files. Audio is carried over LiveKit's infrastructure and recordings (when enabled) are stored under our account.
- Usage stats. Aggregated counts of calls, minutes, and SMS per day, scoped to your user ID. Used to render your dashboard and bill you.
- Operational logs. Cloudflare Workers and Firebase log requests for debugging and abuse prevention. These are retained for 30 days.
What we don't collect
- We do not sell your data. Ever.
- We don't use your call transcripts to train general-purpose AI models.
- We don't track you across other websites or buy data about you from third parties.
Where it lives
- Google Firestore (project
gnbly0) — agent configs, scripts, jobs, contacts, daily stats. Hosted in Google's US multi-region. - Cloudflare Workers — runs the application code. No persistent data is stored on Workers itself.
- LiveKit — runs the real-time audio infrastructure that connects you, our voice agent, and the phone network. See LiveKit's privacy policy for their handling.
- Vonage — telephony carrier. Phone calls traverse their SIP trunk to reach the public phone network.
- Firebase Authentication — your email and sign-in events.
Subprocessors
We share your data only with the providers required to run the service:
- Google (Firebase Auth, Firestore, Cloud Storage)
- Cloudflare (Workers, asset hosting)
- LiveKit (real-time audio infrastructure)
- Vonage (telephony carrier)
- The LLM provider you select per agent (Anthropic, OpenAI, or Groq) — your scripts and call context are sent to whichever you pick
- The TTS provider you select per voice (ElevenLabs, Azure, or OpenAI)
Your rights
- Export everything we have on you. Email us and we'll send a JSON dump.
- Delete your account and all associated data. Email us. Permanent deletion happens within 30 days, except for billing records which we keep for 7 years per US tax law.
- Correct any inaccurate personal info via the Account page or by emailing us.
- Opt out of any non-essential email — we don't send marketing email today, but if we ever do, every message will have an unsubscribe link.
Call recording and consent
Call recording is opt-in per agent. If you enable recording, you are responsible for obtaining consent from anyone you call as required by the laws of your jurisdiction and theirs. Many US states (California, Florida, etc.) require two-party consent. We strongly recommend the agent's first message disclose that the call may be recorded.
Security
Authentication uses Firebase ID tokens verified server-side, plus an HMAC-signed session cookie. Webhooks (LiveKit, Stripe, our own cron) use signature verification. All transport is HTTPS; we don't accept HTTP traffic in production. We're not SOC 2 certified yet — that's on the roadmap as we onboard larger customers.
Changes
We may update this policy. Material changes will be announced via email to your account. The "Last updated" date at the top reflects the most recent revision.