Security
Security & data handling
Public overview of where Chatline runs, how customer data is handled, what external providers are used, and how business buyers can request review before launch.
Infrastructure
Dedicated EU hosting
Chatline.AI runs on dedicated infrastructure in the European Union via Hetzner. Public and account data stay on operator-controlled infrastructure in the EU.
Traffic is served over HTTPS. Conversation data is encrypted in transit and stored encrypted at rest.
Retention
Clear public-chat lifecycle
Anonymous public conversations are automatically purged after 30 days. Registered operators can delete account data and line data through product workflows or support requests.
Chat content is used to operate the service. We do not train models on your chats.
Payments
Billing stays with Paddle
Paddle acts as merchant of record. Chatline does not store raw payment card details. Billing identity, invoices, tax handling, and refunds flow through Paddle.
Public pricing stays readable on the site so buyers can evaluate the service before starting a conversation or account.
AI providers
Provider boundaries are public
Response generation may use OpenRouter, Anthropic, OpenAI. Only the prompt and response context required to generate a reply is sent to these providers.
The current public provider list is available on the providers page.
Business review
Reasonable openness for buyers
Business customers can request privacy, security, DPA, and vendor review before launch or before moving to a paid deployment.
For privacy requests, security questions, DPA review, or vendor review, contact [email protected]. For sales, procurement, or launch-readiness questions, contact [email protected].
Privacy, security, and DPA review requests are reviewed on business days. Initial acknowledgment target: within 2 business days.
Sales, procurement, and launch-readiness questions are reviewed on business days. Initial reply target: within 1 business day.
We do not publish a staffed phone line. Support and vendor review requests are handled asynchronously by email.
Current provider count
6 public providers disclosed
Hosting, billing, model providers, and production monitoring are disclosed publicly so buyers do not need to infer the stack from the page shell.