Twitter API rate limiting: limits, errors and fixes
Twitter API rate limiting explained: X's 15-minute windows, per-app and per-user limits, the 429 error, the headers to read and how to stop hitting the limit.
Piyushh · November 13, 2026 · 3 min read

Twitter API rate limiting means X caps how many requests you can make in a time window, usually 15 minutes. Go over and the API returns HTTP 429, and you wait for the window to reset. There is no single number, because the limit depends on the endpoint and on how you authenticate. Below: how the limits are built, where to read your own, and how to stop hitting them.
Everything about X below comes from X's rate limits page, checked 2026-11-13. X can change these numbers, so confirm them there before you rely on them.
How the Twitter API rate works
Most of it comes down to three ideas.
Limits count requests in a window, usually 15 minutes. X's page says some endpoints use 24 hours or per second instead.
Limits are set per app and per user, and the page treats the two cases separately. Per-app limits apply when you authenticate with a Bearer Token (app-only). Per-user limits apply with OAuth 1.0a or OAuth 2.0 user tokens. Each endpoint has its own number for each case. If you are not sure which one you are using, read our guide on Twitter API authentication.
Each endpoint keeps its own counter, so using up search does not use up tweet lookup.
Twitter API request limits for common endpoints
X's page gives these examples, per 15 minutes:
| Endpoint | Per app | Per user |
|---|---|---|
Recent search, /2/tweets/search/recent | 450 | 300 |
Tweet lookup, /2/tweets | 3,500 | 5,000 |
User timeline, /2/users/:id/tweets | 10,000 | 900 |
Other endpoints have their own docs pages. Check the page for the endpoint you call rather than a blog post, this one included.
The 429 error and the headers
X's page says that when you go over a limit you get a 429 with error code 88 and the message "Rate limit exceeded". Three response headers tell you where you stand:
x-rate-limit-limit: the maximum requests allowedx-rate-limit-remaining: requests left in the windowx-rate-limit-reset: a Unix timestamp for when the window resets
Read them on every response, not only on failures, so you slow down before the 429 instead of after it.
If you're dealing with the error itself, our post on the X rate limit exceeded error covers the causes. For other X access errors such as 403, see Twitter API restrictions.
A rate limit is not a spending limit
A rate limit and a spending limit are easy to confuse. A rate limit means "too many requests, wait". A spending limit means "no more budget". On X's pay-per-use API, running out of credits is its own failure with its own error, covered in is the X API free. When your script stops, work out which limit you hit before you change any code.
How to stop hitting the limit
- Read the headers and sleep. When
x-rate-limit-remainingreaches 0, sleep untilx-rate-limit-reset. - Back off on a 429. Wait, retry, and double the wait each time, up to a cap.
- Ask for more in each request. Use the largest page size the endpoint allows, so you make fewer calls.
- Cache results. If two parts of your program ask the same question, answer the second one from memory.
- Spread the work out. A queue running at a steady rate does better than a burst followed by a long wait.
- Stop polling things that haven't changed. Check a quiet account less often than a busy one.
Here is a small retry helper in Python:
import time, requests
def get_with_backoff(url, headers, params=None, tries=5):
wait = 2
for _ in range(tries):
r = requests.get(url, headers=headers, params=params, timeout=60)
if r.status_code != 429:
r.raise_for_status()
return r.json()
reset = r.headers.get("x-rate-limit-reset")
if reset:
wait = max(wait, int(reset) - int(time.time()) + 1)
time.sleep(min(wait, 900))
wait *= 2
raise RuntimeError("still rate limited after retries")
It uses X's own reset header when the response includes one, and doubles the wait when it doesn't.
Rate limits on a hosted API
On a hosted API instead of X's own, the limits are set by that API. ChatterAPI is independent and not affiliated with X Corp. It limits requests per second by plan: 10 on Starter, 25 on Growth, 50 on Scale. Failed calls cost no credits. A search call looks like this:
curl -s "https://api.chatterapi.com/v1/x/tweets/search?q=rate%20limit&sort=latest" \
-H "Authorization: Bearer $CHATTERAPI_KEY"
Every response carries a credits object showing what the call cost and what is left, so you can track your budget the way you would track the X headers. The X search endpoint docs list the parameters.
Rate limits still exist. They change from 15-minute windows set by X to per-second limits set by your plan, and you no longer need a separate X developer app or a separate bill.
Check the limits you are on
Our own limits and plan prices are on the pricing page.