Privacy Policy
What This Service Is
CyberDiner is a distributed AI inference marketplace. Consumers submit inference jobs through the web application or API; community providers (hardware operators running the cookd daemon) execute those jobs. CyberDiner brokers job dispatch and holds account and usage metadata; it does not own or operate the hardware that runs your jobs.
Data We Collect
Account data (via Clerk). When you create an account, Clerk — our authentication provider — collects your email address and any OAuth identity data you authorize (e.g., GitHub profile if you sign in via GitHub OAuth). We receive your Clerk user ID and basic profile information from Clerk. Clerk's own [privacy policy — link TBD] governs how they handle your authentication data.
Inference content. Your prompts and responses pass through our broker infrastructure (Apache Pulsar) to community providers for computation. We do not persistently log prompt or response content on our own systems.
Provider visibility disclosure. Your inference jobs are executed on hardware owned and operated by community providers. The provider whose hardware executes your job has technical access to the plaintext content of your prompt and response at the time of computation — this is the honest reality of inference on a distributed network of third-party hardware. Providers agree in our Terms of Service not to access, log, or retain content beyond what is required for inference, but this is a contractual and policy control, not a technical guarantee. Do not submit sensitive personal information, credentials, confidential business data, or any content you would not want a technically-sophisticated third party to read.
Usage metadata. We collect job metadata: timestamps, model selected, tier, job duration, completion status, and error codes. This metadata does not include prompt or response content. It is used for metering, operations, and (in future) billing.
Infrastructure logs. Standard server-side logs (IP addresses, request paths, HTTP status codes, error traces) are collected by our hosting infrastructure. [Retention period TBD — owner to specify.]
How We Use Data
- Account data: to authenticate you, enforce access controls, and (in future) associate usage with billing.
- Usage metadata: to measure service usage, detect abuse, produce metering records, and (in future) support billing.
- Infrastructure logs: for debugging, incident response, and security monitoring.
We do not use your data to train AI models, sell advertising, or profile you for third-party marketing.
Data Sharing
We do not sell your personal data. We share data only with:
- Clerk — authentication provider; processes your account and login data per their privacy policy.
- Community providers — your inference job content is dispatched to provider hardware for computation. See the provider visibility disclosure above.
- Hosting infrastructure [provider TBD] — infrastructure-level logs and service data.
- Legal and regulatory: when required by applicable law, valid legal process, or to protect the rights and safety of users and the service.
Alpha Status
CyberDiner is currently in closed alpha. Access is invite-only. Data practices, retention policies, and this privacy policy will be formalized and reviewed by qualified legal counsel before any public launch or billing activation. This document is a draft for initial alpha disclosure purposes.
[Jurisdiction — owner to specify.]
Contact
Privacy questions: feedback@cyberdiner.app