Sub-Processors

Version: 1.0 (DRAFT) Effective date: May 15, 2026 Last updated: April 14, 2026

The following third-party services process data on behalf of ɳSelf for users of our hosted services (cloud.nself.org, task.nself.org, claw.nself.org).

Current Sub-Processors

Sub-processor	Purpose	Data Processed	Location	Transfer Mechanism
Hetzner Online GmbH	Server hosting, compute, storage	All customer data	Falkenstein, Germany	Intra-EEA (none needed)
Vercel Inc.	Frontend hosting, edge rendering	Anonymous web analytics, page requests	USA + global edge	SCCs Module 2
Cloudflare Inc.	DNS, WAF, DDoS protection, R2 storage	Metadata, IP addresses, HTTP requests	USA + global edge	SCCs Module 2 + DPF
Stripe Inc.	Payment processing, fraud detection	Email, payment method, billing address	USA	SCCs Module 2 + DPF
Elastic Email (Emaillabs Sp. z o.o.)	Transactional and marketing email	Email addresses, message bodies	Krakow, Poland (EU)	Intra-EEA (none needed)
OpenAI, LLC	LLM inference for ɳClaw (opt-in)	Prompts, responses	USA	SCCs Module 2 + zero-retention API
Anthropic PBC	LLM inference for ɳClaw (opt-in)	Prompts, responses	USA	SCCs Module 2 + zero-retention API
Groq Inc.	LLM inference for ɳClaw (opt-in)	Prompts, responses	USA	SCCs Module 2
Google Cloud Platform	Optional LLM + email fallback	Varies by feature	USA + EU	SCCs Module 2 + DPF

Changes to This List

We notify cloud.nself.org users by email at least 30 days before adding or replacing a sub-processor. The notification includes the sub-processor’s name, location, processing purpose, and transfer mechanism.

You may object to a new sub-processor within 30 days of notification by emailing privacy@nself.org with a reasonable justification. If we cannot accommodate the objection, either party may terminate the affected services with 30 days’ notice.

Self-Hosted Users

If you run ɳSelf on your own infrastructure, none of these sub-processors apply. Your data stays on your servers. The only external call the CLI makes (if telemetry is enabled) is to ping.nself.org for version checks and anonymous telemetry. You can disable this entirely with nself config set telemetry false.

AI Model Providers (Detail)

OpenAI, Anthropic, and Groq are used only for hosted ɳClaw and AI features. They are opt-in: you choose which providers to enable in your ɳClaw settings. You can disable all third-party providers and use only self-hosted models (Llama via Ollama), in which case your prompts never leave our Hetzner servers.

All third-party AI providers are configured with zero-retention / no-training API settings where available. Your prompts and responses are not used to train any model.

Contact

Questions about our sub-processors: privacy@nself.org