Privacy policy.

This policy covers tegan.ai, app.tegan.ai, meet.tegan.ai, and any embedded Tegan agent running on a customer site. “You” means either a Tegan customer (the company running an agent) or an end user (the person talking to a Tegan agent). Where the distinction matters, we'll call it out.

From customers:

• Account info — name, work email, company name. Used to sign you in and bill you.
• Knowledge sources you connect — URLs to crawl, YouTube transcripts, pasted notes. Stored to power your agent.
• Configuration — agent personality, goals, opening prompts, team-member access.

From end users talking to an agent:

• Voice audio for the duration of the call. Used to power the conversation and stored as a recording for the customer to review.
• Screen video, only if screen-share is enabled and the user agrees to share. Used for in-call vision and stored alongside the audio.
• Full transcript with timestamps and turn-taking metadata.
• Email address, only if the customer has enabled the email-gate. Stored against the session.
• Basic browser metadata — device type, locale, IP-derived region. Used for compatibility and quality investigation.

• We do not record or transcribe anything outside an active Tegan session. The mic only opens when the user starts a conversation, and it closes when they end it.
• We do not access your screen except during an active call where the user has explicitly shared their screen.
• We do not sell personal data to third parties. We don't have an ad business.
• We do not train shared models on your customer data. Each customer's knowledge base, transcripts, and configuration are scoped to that customer.

Customer data is used to operate the agent you configured: crawl your knowledge sources, run the voice conversation, store and surface the recording, score the session, and show analytics in your dashboard. Aggregated, de-identified usage data (e.g., “average call length across all agents”) may be used to improve the product.

End-user data captured during a session is shown to the customer whose agent the user talked to. End users see a clear notice that the call is recorded before the conversation starts.

We use a small set of vendors to deliver the service — infrastructure (Cloudflare), voice models, transcription, and observability. The current sub-processor list is available on request to customers. Changes to material sub-processors are communicated in advance.

Session recordings, transcripts, and scores are retained for as long as the customer's workspace is active. Customers can delete individual sessions, or request a full workspace wipe at any time — write to privacy@tegan.ai.

End users who want their conversations deleted should contact the customer who runs the agent they talked to (they control the workspace). If you can't reach the customer, write to us and we'll route it.

Data in transit is TLS-encrypted end-to-end. Data at rest is encrypted on managed cloud storage. Access to production data is limited to a small set of engineers with audited logins. A SOC 2 program is on the roadmap; see the security page for current posture.

If GDPR, UK GDPR, or CCPA applies to you, you have rights to access, correct, port, or delete your personal data. Customers can act on these requests via the dashboard for their own users; end users can contact privacy@tegan.ai.

Tegan is built for B2B SaaS contexts. The service is not directed at children under 16 and we don't knowingly collect data from them. If you believe we have, let us know and we'll delete it.

Material changes to this policy are announced via the dashboard for active customers. The “Last updated” date at the top of this page always reflects the current version.

Privacy questions: privacy@tegan.ai. General contact: hello@tegan.ai.