TokenResets
한국어

Sources

Every published event links to the original source. This page describes what is monitored and how much weight each kind of source carries.

Coverage by provider

ProviderMonitoring sinceEventsSource status
CODEXOpenAI Codex2025-09-1761Fresh
CLAUDEAnthropic Claude2026-04-1620Delayed

Monitored source registry

These are the enabled public endpoints, accounts, and policy references used by the current build. The list is generated from the same provider registry that configures detection; it is not maintained as a separate marketing list.

Provider registry 3.3.0 · detection registry 1.3.0

Codex

OpenAI Codex

Inspect published evidence
Public sourceRoleTrustHow it is used
Codex changelog feed (opens in a new window)developers.openai.com/codex/changelog/rss.xmlOfficial announcement feedS1Independent product announcements. Full feed text and original dates are reverified; generic reset instructions are not new grants.
OpenAI news feed (opens in a new window)openai.com/news/rss.xmlOfficial announcement feedS1Independent product announcements. Full feed text and original dates are reverified; generic reset instructions are not new grants.
Codex status history RSS (opens in a new window)status.openai.com/history.rssOfficial status feedS0Status and incident corroboration. Each item is checked for reset relevance.
@thsottiaux on X (opens in a new window)x.com/thsottiauxAnnouncement accountS2Codex product lead. Direct reset statements can become event evidence.
Codex pricing and limits (opens in a new window)developers.openai.com/codex/pricingPolicy referenceS1Policy, plan, product, and affected-scope verification.
Codex plan help (opens in a new window)help.openai.com/en/articles/11369540-using-codex-with-your-chatgpt-planPolicy referenceS1Policy, plan, product, and affected-scope verification.
Codex changelog (opens in a new window)developers.openai.com/codex/changelogPolicy referenceS1Policy, plan, product, and affected-scope verification.
Codex changelog feed (opens in a new window)developers.openai.com/codex/changelog/rss.xmlPolicy referenceS1Policy, plan, product, and affected-scope verification.
OpenAI news feed (opens in a new window)openai.com/news/rss.xmlPolicy referenceS1Policy, plan, product, and affected-scope verification.
OpenAI Status (opens in a new window)status.openai.comPolicy referenceS0Policy, plan, product, and affected-scope verification.
codex-resets.com data feed (opens in a new window)codex-resets.com/api/resetsEnumeration catalogS3 · ReportedAllowlisted structured ID, timestamp, and reset type may publish as reported evidence. Feed prose and statistics are not copied.
claude-resets.com data feed (opens in a new window)claude-resets.com/data/resets.jsonEnumeration catalogDiscovery onlyCandidate enumeration only. Original evidence is checked before publication.

Claude

Anthropic Claude

Inspect published evidence
Public sourceRoleTrustHow it is used
Claude status history RSS (opens in a new window)status.claude.com/history.rssOfficial status feedS0Status and incident corroboration. Each item is checked for reset relevance.
@ClaudeDevs on X (opens in a new window)x.com/ClaudeDevsAnnouncement accountS1Anthropic official developer account. Direct reset statements can become event evidence.
@lydiahallie on X (opens in a new window)x.com/lydiahallieAnnouncement accountS2Claude Code staff. Direct reset statements can become event evidence.
Understanding usage and length limits (opens in a new window)support.anthropic.com/en/articles/11647753-understanding-usage-and-length-limitsPolicy referenceS1Policy, plan, product, and affected-scope verification.
Usage limit best practices (opens in a new window)support.anthropic.com/en/articles/9797557-usage-limit-best-practicesPolicy referenceS1Policy, plan, product, and affected-scope verification.
Claude Code plan usage (opens in a new window)support.anthropic.com/en/articles/11145838-using-claude-code-with-your-pro-or-max-planPolicy referenceS1Policy, plan, product, and affected-scope verification.
Anthropic Status (opens in a new window)status.claude.comPolicy referenceS0Policy, plan, product, and affected-scope verification.
Anthropic news (opens in a new window)www.anthropic.com/newsPolicy referenceS1Policy, plan, product, and affected-scope verification.
claude-resets.com data feed (opens in a new window)claude-resets.com/data/resets.jsonEnumeration catalogDiscovery onlyCandidate enumeration only. Original evidence is checked before publication.

From detection to publication

  1. Enumerate. Enabled accounts, feeds, and catalogs are checked every 10 minutes.
  2. Verify. Deterministic rules evaluate the original statement, identity, timing, and affected scope.
  3. Publish or reject. Only qualifying evidence enters the public ledger; rejected candidates do not become events.

An allowlisted catalog fact can publish only as Reported and is excluded from verified-only interval statistics. Other catalogs remain discovery-only. Event pages preserve the public link supporting each record.

Source trust levels

S0
Official machine-readable status or API.
S1
Official documentation, help centre, changelog, or an official company account.
S2
An approved staff account making a direct statement about their own product.
S3
Credible secondary reporting. Requires corroboration.
S4
Community reports. Supporting evidence only, never sufficient to publish.

What we do not do

  • We do not ask for, store, or transmit your provider account credentials.
  • We do not bypass logins, CAPTCHAs, or access restrictions.
  • We do not republish source content in full. We link to the original.
  • When a source is removed, we stop showing the quotation and re-evaluate the event's confidence rather than silently keeping the claim.

When sources go quiet

If monitoring is interrupted, the affected provider is marked as delayed or stale and the last successful check is shown. Stored history stays available. We never present "no reset happened" as a fresh conclusion during an outage.