Transparent by design
Source policy
How public sources are collected, analyzed, retained and displayed.
GitHub
We use authenticated public REST search and configured repository release endpoints, respecting API limits. Collection is bounded to configured tools and time windows. Raw issue/discussion-like content expires after 90 days. External usernames are replaced with source-specific HMAC fingerprints in analytical records.
Hacker News
We use the HN Algolia search API for recent matching stories and comments. Adapter tests use saved payloads. Collection and commercialization assumptions must be reviewed against current API and content policies before public launch.
Official context
Configured official repository releases are stored as contextual events. Pricing is only authoritative after official verification. Release timing can be correlated with reported changes but is never presented as proof of causation.
Public display and retention
We display source links, timestamps, short generated analytical summaries and aggregates. We do not republish full comments. Raw content and observations expire after the configured retention period; historical aggregate snapshots remain indefinitely.
Reddit access
An optional OAuth adapter can collect recent posts and comments from configured Reddit communities. It remains disabled until permitted API access and credentials are configured. Web-search excerpts are not treated as verified comment records. Removed content and known bot entries are excluded; requests and collection are bounded.
Collection cadence
Collection is configured once daily with a 26-hour overlapping lookback and a 200-candidate total daily cap across tools and sources. Daily snapshots use stored observations; a weekly review refreshes on Mondays without additional AI calls.
Excluded sources
X scraping and arbitrary-site scraping are not included. Every new commercial data source requires review of collection, storage, analysis and derived-display permissions.