What Snashd stores

Stored

• Headline and dek (first ~320 characters) per story
• Canonical URL, outlet id, language
• Publish and capture timestamps
• Headline edits, kept as a revision history — never silently overwritten

Never stored

• Article bodies, images, or paywalled content — the service links out to the publisher
• User browsing history. There are no user accounts and no third-party analytics.

Retention

60 days, enforced on every load; older records are pruned automatically.

Fetching

Polite RSS fetching only: robots.txt is honoured and disallowed feeds are refused, requests carry a descriptive user agent, and each host is rate-spaced.

Prototype status

This build runs on localhost only. Nothing leaves the machine except the outbound RSS fetches themselves.