Trust Center

Subprocessors

Autark routes API requests to third-party LLM providers. Your prompts are processed in memory and never stored, but they transit through these providers. Each provider's own data handling policies apply during that transit.

ProviderServiceData ProcessedLocationTrust Page
GroqLPU inference for the Flash tierPrompt payloads (in-memory, not stored)US (California)View
Fireworks AIInference for Pro and Ultra tiers, including DeepSeek modelsPrompt payloads (in-memory, not stored)US + SingaporeView
Nebius AIInference for select models (coming soon)Prompt payloads (in-memory, not stored)Finland + EUView

A Data Processing Agreement (DPA) covering these subprocessors, including EU Standard Contractual Clauses, is available on request — contact us.