Subprocessor Register

01Current subprocessors

• Cloud infrastructure (GCP) — compute, storage, managed Postgres. Data: all tenant data (encrypted). Region: customer-selected.
• Anthropic — LLM inference via the Router. Data: prompt + context at run time. Region: US / EU.
• Google (Gemini) — LLM inference via the Router. Data: prompt + context at run time. Region: US / EU.
• OpenAI — optional LLM inference via the Router. Data: prompt + context at run time. Region: US.
• Object storage / CDN — artifact and asset delivery. Data: generated artifacts (encrypted). Region: customer-selected.
• Error & ops telemetry — reliability monitoring. Data: operational metadata only, no payloads. Region: US / EU.

02Notes

Model subprocessors are inference-only and contractually barred from training on customer data. Bring-your-own-key tenants call providers under their own agreements; air-gapped deployments use local models and engage no model subprocessor at all.

To subscribe to change notices, email privacy@cortexaios.com with subject "Subprocessor notices".