Sub-Processors
We use a limited set of third-party service providers to operate the Services. Each has contractual confidentiality and data-protection obligations. For full details, see our Privacy Policy.
Infrastructure
| Sub-Processor | Purpose | Data Shared | Location |
|---|---|---|---|
| Supabase (via AWS) | Database hosting, authentication, storage | Account, Workspace, encrypted API keys, Request Metadata, Billing | United States |
| Cloudflare | Edge routing, request caching, KV storage, DDoS protection | Request Metadata, API key hashes, rate-limit state | Global (edge network) |
| Stripe | Payment processing | Billing Information, email | United States |
| Railway | Backend API hosting | Account, Request Metadata, Billing | United States |
| Vercel | Frontend hosting | Session cookies, IP address | United States |
| Upstash | Async job queues (billing events, metrics ingestion) | Request Metadata, credit transactions | United States |
| Resend | Transactional email delivery | Email address, message content | United States |
Inference providers
When a request is routed to a model served by one of these providers, that provider may process the request for inference.
- Purpose
- LLM inference and model routing
- Data processed
- Content submitted for inference, generated outputs, and request metadata necessary to route and deliver inference requests.
- Location basis
- Provider location identifies the provider legal-entity or service location based on official sources and Auriko's applicable account context. It does not mean inference data is processed or stored only in that country; processing and storage are governed by provider terms, data-processing agreements, account settings, and service configuration.
| Sub-Processor | Service | Provider location |
|---|---|---|
| Anthropic | Claude model inference | United States |
| DeepInfra | Serverless inference for supported models | United States |
| DeepSeek | DeepSeek model inference | China |
| Fireworks AI | Serverless inference for supported models | United States |
| Google AI Studio | Gemini model inference | United States |
| MiniMax | MiniMax model inference | China |
| Moonshot AI | Kimi model inference | Singapore |
| OpenAI | OpenAI model inference | United States |
| SiliconFlow | Serverless inference for supported models | Singapore |
| Together AI | Serverless inference for supported models | United States |
| xAI | Grok model inference | United States |
| Z.AI | GLM model inference | Singapore |
We maintain this sub-processor list and update it when changes occur. We communicate material sub-processor changes to business customers per the DPA.