Chips Studio / Rill

AI Services and Your Data

Rill sends directly to your configured service, without a Rill inference proxy. You provide an authorized key, address and exact model. Costs, access and capabilities belong to that service.

Chat requests may contain the current message, full applicable original branch history, system prompt, relevant current/historical attachments and necessary tool protocol state. Group participants have independent context by default. Directory requests contain authentication/configuration, not conversations.

Official information

Retention and training vary by account, plan, region and settings. Check current terms before sending private or regulated information. Directory listings do not certify access, safety or inference success.

Custom and local destinations

You choose OpenAI-compatible endpoints. Confirm the operator, local/remote address, authentication and privacy policy. Plain HTTP has no transport encryption. Local servers can themselves send data elsewhere; Rill cannot promise otherwise.

Rill Demo is an explicitly labeled offline sample, without a key, network or AI understanding. Rill 1.0 disables executable HTML preview, voice and cloud sync.

Privacy · Support