Privacy and beta measurement
Free beta, no sign-in. CSV parsing, naming rules, findings and reports run in your browser. We do not upload campaign URLs, CSV contents, filenames, rules or reports. There is no browser storage, account, billing or saved history. Reloading clears the working state; downloaded files remain under your control until you delete them.
Optional, best-effort aggregate counters
The tool page sends only two fixed enums to same-origin POST /telemetry: event (visit, check_completed, html_download_requested or csv_download_requested) and acquisition (direct, search or other). Acquisition uses a finite hostname-only search-referrer allowlist, not query parameters. No referrer text is sent. Article views are not counted. Blocking this endpoint, being offline or an analytics error does not prevent checking or downloading. No retries or offline queue.
Requests omit credentials and referrers. We do not set cookies or persistent identifiers, store individual events or retain IPs in our application database. The database contains only UTC day, event enum, acquisition enum and count (at most 12 buckets per day). Downloads count an initiated request, not a confirmed save. Completed checks require at least one analyzed row; mixed skipped rows can still count. Counts are events, not unique people, qualified leads or causal article attribution. Repeats, bots, offline losses and caps can bias them.
Retention, limits and providers
Daily scheduled cleanup keeps the current UTC day and previous 29 days of aggregates (30 daily buckets); deletion runs daily, so an expired bucket can remain until the next successful cleanup. Studio-lead owns deletion and cleanup monitoring. The global ceiling is 5000 accepted events per UTC day, enforced atomically; an early 100 requests/minute/isolate gate is not a global rate limit. Rejected requests still consume provider compute: these are not a hard spending cap. There is no public metrics read endpoint.
Cloudflare delivers this site and processes connection metadata such as IP addresses and request information as a service provider. This is distinct from our aggregate counter database. Application code does not log request bodies, URLs, headers or referrers; Worker observability and logpush are configured off and must be verified by the release executor. We do not promise that Cloudflare never processes connection metadata. Cloudflare privacy policy.
Local report safety and limitations
Inputs are limited to UTF-8 .csv, 262144 bytes and 50 data rows. No server file retention exists because files are not uploaded. Downloaded reports intentionally contain your supplied campaign values; handle and delete them appropriately. HTML reports are self-contained, contain no telemetry and make no HTTP requests offline. Destinations are not fetched or URL-validated, redirects are not checked, and a report is not proof that links work. A skipped row is not approved.