There is no universal API quota reset time. The answer depends on the provider, the quota dimension, and the account or project that owns the limit. A 429 caused by requests per minute may clear after a rolling countdown or at the next synchronized interval. A daily quota may reset at a provider-defined midnight, while a monthly spend cap, exhausted prepaid balance, or approved usage limit may require a billing or limit change instead of waiting.
Start by reading the complete error response and its headers. Look for Retry-After, provider-specific reset fields, the quota name, and the scope (key, project, organization, or resource). Only then convert the documented reset time to your own timezone.
“Quota” can mean several different limits
API products commonly enforce more than one control at the same time. Exceeding one does not imply that the others are exhausted.
| Quota or limit | What it measures | Typical remedy |
|---|---|---|
| Request rate (RPM) | Requests in a minute or another short window | Honor the retry countdown, then throttle or queue work |
| Token rate (TPM) | Input and output tokens processed in a short window | Wait for token capacity, reduce prompt size, or lower concurrency |
| Daily requests (RPD) | Calls allowed during a provider-defined day | Wait for the documented daily boundary or change the project tier |
| Monthly usage or approved limit | Accumulated usage authorized for an organization | Wait for the monthly cycle if applicable, or request a higher limit |
| Spend limit | Dollar cap configured for an organization or project | Raise or remove the cap, or wait for its monthly reset |
| Prepaid balance or credits | Money available to pay for requests | Add credits; retrying alone does not restore access |
Therefore, “I got a 429” is not enough information to predict a reset. A temporary rate error and an insufficient_quota or billing error can look similar to an application that only logs the status code.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →#1 Best Overall
How to find the actual reset time
- Capture the full response. Record the HTTP status, JSON error type and code, response headers, request ID, and the endpoint or model used.
- Identify the dimension. Decide whether the message refers to requests, tokens, daily calls, monthly usage, spend, or credits.
- Read reset metadata before guessing. Prefer
Retry-Afterand provider-specific reset headers over a fixed sleep. - Check scope and account state. A key may inherit a project limit; a project may inherit an organization limit. Open the provider’s limits, billing, and usage pages to confirm the current tier and balance.
- Convert the boundary correctly. A Unix timestamp or a provider timezone must be converted to the timezone used by your scheduler. Daylight-saving changes can make a “midnight” assumption wrong.
Inspect headers with cURL
curl -i https://api.example.com/v1/resource
-H "Authorization: Bearer $API_KEY"Keep the headers in logs, but redact authorization tokens and personal data. A useful log record includes the remaining requests and tokens, reset values, Retry-After, status, endpoint, project, and model.
Use adaptive backoff, not a fixed midnight retry
For a temporary rate error, pause for the server-provided delay. If no delay is supplied, use exponential backoff with jitter (for example, 1, 2, 4, 8 seconds with a random offset) and a maximum retry count. Do not retry billing, credit, or hard-spend errors in a tight loop: they will not be repaired by additional requests.
OpenAI API: rate windows are different from quota and billing limits
OpenAI exposes short-term capacity in response fields such as x-ratelimit-remaining-requests, x-ratelimit-remaining-tokens, x-ratelimit-reset-requests, x-ratelimit-reset-tokens, and x-ratelimit-reset-project-tokens. A temporary 429 may also include Retry-After. These values tell you how much capacity remains and how long until the applicable short-term window resets; use them instead of assuming a clock-time boundary.
OpenAI also separates those windows from financial and organizational controls. Its documentation states that “OpenAI sets an approved monthly usage limit for each organization.” Configurable organization and project spend limits are separate. An insufficient_quota-style response can therefore mean an approved usage limit, a project or organization cap, or exhausted prepaid credits rather than a burst-rate problem.
Rank #2
- Used Book in Good Condition
OpenAI diagnostic sequence
- Read the HTTP status, error type, and error code.
- For a request or token rate error, honor
Retry-Afteror the relevantx-ratelimit-reset-*countdown. - For
credit_balance_exhausted, add prepaid credits before retrying. - For an organization or project spend-limit error, review permissions and the applicable limit. Waiting for the next monthly cycle may be necessary when the cap is intentionally enforced.
- Open the Limits page to verify the organization tier and approved monthly usage limit.
Documented example tiers include Free and Tier 1 at $100 per month, Tier 2 at $500, Tier 3 at $1,000, Tier 4 at $5,000, and Tier 5 at $200,000. These are example tier values and can change; they are not a promise that every account has those limits.
Gemini API: RPM, TPM and RPD reset independently
Google documents three independent Gemini dimensions: requests per minute (RPM), tokens per minute (TPM), and requests per day (RPD). You can have remaining token capacity while requests per minute is exhausted, or vice versa. Diagnose the dimension named in the error or quota metrics rather than waiting for all three to reset.
The documented RPD boundary is midnight Pacific Time. That is a daily boundary, not “24 hours after my last request.” Gemini limits are applied per project, not per API key, so replacing a key does not create a fresh project allowance.
Google Cloud APIs: intervals are service-specific
Google Cloud does not provide one reset rule for every API. Each service defines its quota interval. Compute Engine gives a concrete example of synchronized one-minute intervals: if a project reaches its maximum at 10:00:15, capacity can return at the next boundary, such as 10:01:00, rather than exactly 60 seconds after the request.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minuteThis distinction matters for schedulers. A worker that sleeps 60 seconds after every 10:00:15 failure can wake at 10:01:15 and lose throughput, while a worker that retries immediately at 10:01:00 can succeed. Use the service’s documented interval and any returned retry metadata.
GitHub API: read the resource-specific timestamp
GitHub’s REST rate-limit endpoint returns a reset Unix timestamp for each resource. Convert that timestamp to the timezone used by your operations dashboard and wait until that resource’s boundary. REST and GraphQL have separate rate-limit systems, so a reset for one API family does not automatically describe the other.
Why waiting did not fix the error
- Wrong dimension: You waited for RPM, but TPM, RPD, or a monthly limit is exhausted.
- Wrong scope: The key is fine, but the project or organization has reached its cap.
- Billing state: Prepaid credits are depleted or a payment issue has suspended usage.
- Hard spend cap: A configured organization or project limit requires a change or the next monthly cycle.
- Clock conversion: You treated Pacific midnight, a Unix timestamp, or a synchronized interval as local midnight.
- Independent services: A REST resource, GraphQL resource, model, or region has its own bucket.
Do not “solve” these cases by rotating API keys. If the quota is project-, organization-, or resource-scoped, new keys inherit the same state.
Designing clients that recover safely
Separate retryable and non-retryable failures
- Retry a temporary rate response when the provider supplies a delay or reset countdown.
- Queue work and lower concurrency when remaining request or token capacity is nearly zero.
- Stop and alert on credit, payment, approved-usage, or hard-spend errors.
- Keep idempotency in mind: a timed-out request may have completed, so avoid duplicating non-idempotent operations.
Make quota state observable
Emit metrics for remaining requests, remaining tokens, reset timestamps, 429 counts, billing errors, and queue age. Tag them by provider, project, organization, model, API family, and region where those fields exist. Alert before exhaustion so a batch can be deferred instead of failing at the boundary.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Plan for synchronized boundaries
For interval-based quotas, add jitter to large job batches and spread requests across the window. For daily quotas, calculate the provider’s boundary in its documented timezone. For monthly limits, forecast usage and leave a safety margin; a monthly reset is not a substitute for capacity planning.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Or skip the browser setup
If you need a clean, repeatable image of a provider’s limits or usage page for an incident record, ScreenshotNeo is a website screenshot API and MCP server. It accepts consent banners like a visitor, removes more than 60 known consent platforms plus newsletter popups and chat widgets, and bills only clean shots: bot checks, CAPTCHAs, blank pages, timeouts, failed loads, and cache hits cost nothing. Responses identify the page verdict and billing result in X-Page-Verdict and X-Billed headers. Its MCP tools—take_screenshot, get_page_info, and capture_pdf—let Claude, Cursor, or another MCP client capture pages without your own browser automation.
One GET request is enough:
curl -G "https://api.screenshotneo.com/v1/shot" -d access_key=YOUR_API_KEY --data-urlencode url=https://screenshotneo.com/docs/ -o shot.webpSee the complete parameter reference in the ScreenshotNeo documentation. The same request in Python:
import requests
r = requests.get("https://api.screenshotneo.com/v1/shot", params={"access_key": "YOUR_API_KEY", "url": "https://screenshotneo.com/docs/"}, timeout=90)
open("shot.webp", "wb").write(r.content)And in Node.js:
const q = new URLSearchParams({ access_key: 'YOUR_API_KEY', url: 'https://screenshotneo.com/docs/' });
const res = await fetch(`https://api.screenshotneo.com/v1/shot?${q}`);Every feature is included on every plan. The Free plan provides 1,000 screenshots each month with no card; paid plans start at $5 for 3,000. Sign up free for ScreenshotNeo.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Quick decision checklist
- Is the error a temporary rate response or a billing/quota response?
- Which dimension is named: requests, tokens, daily calls, monthly usage, spend, or credits?
- What do
Retry-Afterand reset headers say? - Is the limit owned by a key, project, organization, or resource?
- What timezone or synchronized interval does the provider document?
- Would throttling, adding credits, raising a limit, or contacting support fix it faster than waiting?
Frequently Asked Questions
Does every API reset at midnight?
No. Midnight applies only to providers and quota dimensions that define a daily boundary. Other limits use rolling windows, synchronized intervals, or monthly cycles.
Best Value
How long should I wait after a 429?
Use the response’s Retry-After or reset countdown. If neither is present, apply bounded exponential backoff with jitter and stop after a defined number of attempts.
Can creating a new API key reset my quota?
Usually not when the quota is scoped to a project, organization, or resource. Check the scope shown by the provider before rotating keys.
Why does an insufficient-quota error remain after the rate window reset?
It may represent exhausted credits, an approved monthly usage limit, or an organization/project spend cap. Those conditions require a billing or limit change, or the next monthly cycle.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
