# smry public reader privacy

The r.smry.ai reader processes the public source URL supplied by the caller so
it can retrieve or locate a cached extraction and return cleaned text. The
Worker applies per-IP abuse limits, but its control plane stores only a SHA-256
digest for coordinated rate enforcement rather than the raw IP address. Usage
reporting is aggregate and groups requests by a finite agent-family category,
status, request mode, cache outcome, and cold-miss outcome. It does not retain
raw User-Agent strings, target URLs, or article bodies in those usage rollups.

The public reader does not need a smry account or API key. Do not send private
URLs, cookies, authorization headers, subscriber credentials, or personal
documents to it. Account-owned library and feed tools use separate authenticated
interfaces on api.smry.ai. The full smry privacy policy, including account and
consumer application data practices, is canonical at https://smry.ai/privacy.

Security and privacy questions can be sent to michael@smry.ai without including
secrets or private source content.