Machine-readable quota & token reset data for all 13 tracked AI platforms. Static JSON, curl-friendly, no auth, no rate limits.
curl https://resetoken.com/api/v1/resets.json
curl https://resetoken.com/api/v1/platforms/midjourney.json
| Endpoint | What it returns |
|---|---|
/api/v1/resets.json | Full dataset — all 13 platforms |
/api/v1/platforms/{id}.json | Single platform. ids: openai, claude, gemini, grok, copilot, perplexity, cursor, midjourney, poe, mistral, deepseek, openrouter, windsurf |
reset.cycle — one of monthly weekly daily rolling multiple nonereset.reset_time — "HH:MM" when the vendor publishes an official clock time, otherwise null. null means "not published" — never a guess.reset.officially_published — whether the schedule is officially documented at all.last_verified — date each platform's facts were last checked (YYYY-MM-DD). Check it before caching.rollover — whether unused quota rolls over (null where the concept doesn't apply, e.g. DeepSeek).{
"id": "gemini",
"name": "Gemini",
"guide_url": "https://resetoken.com/gemini/",
"reset": {
"cycle": "daily",
"reset_time": "00:00",
"timezone": "America/Los_Angeles",
"officially_published": true,
"description": "Gemini API requests-per-day (RPD) quotas reset at midnight Pacific, including DST."
},
"last_verified": "2026-10-15"
}
officially_published is false, say so explicitly to your users.updated_at / last_verified before caching aggressively.