Privacy and data handling
A precise summary of identity, keys, prompts, and operational data.
Privacy and data handling
This page describes the intended service behavior. See the privacy policy for the product-facing notice and terms for use conditions.
Anonymous requests
Anonymous limits use the Cloudflare CF-Connecting-IP value only to derive a
daily rotating HMAC-SHA-256 subject. The raw IP is not written to D1, KV,
analytics, logs, errors, or responses. The subject changes at UTC midnight and
is not intended to track a person across days.
Signed-in users
GitHub sign-in stores the minimum identity needed to operate the account: numeric GitHub ID, login, optional avatar URL, account status, and timestamps. The OAuth access token is used during the callback and discarded. It is not stored in D1, KV, a cookie, a log, or an API response.
API keys
The raw llmfaucet_... value is shown once when created. The service stores
only a SHA-256 hash and a non-secret prefix. Existing keys cannot be recovered;
they can be revoked and replaced.
Content and analytics
Prompt and completion content is not retained by default. Operational metadata may include route, status, latency, request timestamp, and aggregate usage needed to run the service. Do not send secrets, private credentials, or sensitive source code to a shared preview service.
Waitlist and early testers
Applications store qualification details such as tool, use case, operating system, expected volume, tester goal, and agreement timestamps. Reviewer notes are private. Waitlist records and events are removed with account deletion. Sponsorship is not required for selection.
Your controls
- Revoke individual API keys from the dashboard.
- Sign out to revoke the current session.
- Delete the account to remove personal records and active credentials.
- Contact the project through GitHub Issues for privacy questions.
Aggregate, non-identifying service statistics may remain after deletion. See early testers for preview-specific handling.