Your data flows the way you choose

When you run Cline in your own VPC, inference requests flow directly from your environment to your chosen model providers — Cline doesn't see your code or prompts. When you use Cline-managed inference, we see requests only to route them to the model, and your code is never used to train models.