What it is
A free header reader for rate limits. It understands Retry After, the X RateLimit family, the IETF RateLimit fields and the formats used by GitHub, OpenAI, Anthropic, Shopify and Discord.
429 explained in plain words
Your API said 429. Here is what it actually meant.
Paste the response headers. Rate Limit Story reads the limit, what is left, when it resets and how long to wait, then writes a four panel story and a ready reply for the client who hit the wall.
- Client
Sending batch 501 of 600…
- API
Too many requests! Token bucket is empty.
- Rate Limit Story
Wait 20 s, then retry with a little jitter.
Who it is for
API owners, developer support, partner managers and anyone whose client just sent a screenshot of a 429 with no context.
What it solves
Rate limit headers are cryptic, mix Unix times with seconds and differ per vendor. People guess, retry too fast and get blocked for longer.
Tell the story of a response
Everything is read in this tab. Nothing you paste is uploaded or stored.
Load a sample
Paste raw headers, curl -i output or a JSON object of headers. The status line is optional.No story yet
Paste the headers from a failing call, or load one of the samples.
Headers never leave this page. Remove auth or cookie headers before sharing a screenshot anyway.
Headers it understands
Mix headers from different styles in one paste. Names are not case sensitive.
Retry After
Retry-After: 20Seconds or an HTTP date. When present, it wins over every other reset hint.
X RateLimit family
X-RateLimit-Limit: 5000
X-RateLimit-Remaining: 0
X-RateLimit-Reset: 1792661820Used by GitHub, Twitter style APIs and many others. Reset can be Unix seconds, milliseconds or seconds from now, and it says which one it picked.
IETF RateLimit
RateLimit-Policy: "default";q=100;w=60
RateLimit: "default";r=7;t=42The standard fields with quota, window, remaining and time to reset.
OpenAI and Anthropic
x-ratelimit-remaining-tokens: 0
x-ratelimit-reset-tokens: 18.4sSeparate request and token buckets with durations like 6m0s. The empty bucket is marked as the one holding you back.
Shopify and Discord
X-Shopify-Shop-Api-Call-Limit: 38/40
X-RateLimit-Reset-After: 2.75Used out of size counters, reset after seconds, buckets, scopes and global limits.
How to use it
Copy the headers
From the browser network tab, curl -i, your logs or the screenshot a client sent. Only the header lines matter.
Paste and tell the story
Each bucket is read on its own and the one that is empty or closest to empty becomes the main character.
Read the four panels
What happened, where you stand, when it resets and what to do, with times in UTC and in your own time zone.
Send the reply
Copy the ready reply into the ticket, or copy the full story for your own notes.
Good to know
- It explains the headers you paste. It does not call the API and cannot see limits the API keeps to itself.
- When a reset value is ambiguous, numbers above one billion are treated as Unix time and smaller ones as seconds.
- The safe pace needs both a limit and a window length. Many APIs only send the limit.
- Times in your zone use the clock of this device.
Questions
Why does GitHub send 403 instead of 429?
GitHub answers some rate limited calls with 403. The story still treats it as limited when remaining is zero.
Which wins, Retry After or the reset time?
Retry After. It is the server telling you exactly how long to wait, while reset describes the whole window.
What is jitter and why add it?
A small random extra delay. Without it, every client that was blocked retries at the same second and trips the limit again.
Is it safe to paste real headers?
The page sends nothing anywhere. Still, remove Authorization and cookie headers before you share anything.