Show the work.
Keep the claim honest.
This report turns the monitor's public record into inspectable quality metrics. It separates collection cadence, evidence coverage, and source-to-observation timing so a fast check is never mistaken for instant delivery or an account-level guarantee.
31 confirmed · 7 early signal. Corroborating records are never silently promoted to confirmation.
Every public event in this report has a human-readable evidence page and a machine-readable receipt.
No qualifying source-to-observation sample is being claimed yet.
No Codex login, credential, private account telemetry, or source address is exposed by this report.
What is being measured
Collection
Independent server-side checks run on a 60-second cadence. Last successful check: Sep 3, 2026, 2:44 PM UTC.
Record quality
56 of 66 records include an observed time; 53 include a stated scope; 38 include an event type. These are completeness facts, not accuracy guarantees.
Timing without marketing math
Only primary events with an explicit source-fetch timingQuality and first-observed timestamp are included. Migration backfills and external ingests are excluded.
Within a 24-hour measurement window.
Historical source-to-observation measurements are not a delivery SLA and do not prove that every account changed.
Why this is a product advantage
Most reset pages ask users to trust a headline, a forecast, or a refresh label. ResetPulse makes the evidence portable: a developer can fetch the verified packet, validate the receipt, embed the badge, or use the one-file client. A competitor can be faster on a particular event; this report makes that claim testable rather than assumed.
