Calculate rate limiter

Convert a request allowance and window into steady rates, average spacing, and token-bucket refill figures.

freeworks offlinenothing uploaded
ToolRate Limiter Calculator
Input
Output
—
Put this on your own site

The frame below runs the same code as this page, in the reader's own browser. Nothing is sent to us, and nothing is sent to you.

Pick a dark background and the text and panels follow it, so the frame stays readable on a dark page.

Preview

How it works

Requests divided by window seconds gives the steady per-second rate and its reciprocal gives average spacing; the same rate refills tokens. Burst size sets bucket capacity and refill time, while the sliding-window lines restate the supplied maximum and duration.

  • One hundred per minute and ten burst tokens are editable demonstration values rather than an API recommendation.
  • Longer projections assume the same limit runs continuously without secondary quotas or downtime.

Worked example

API Rate Limit
Calculate rate limiting for 100 requests per minute with burst of 10
Input
											Requests: 100
Window seconds: 60
Burst size: 10
Output
												=== Rate Limit Calculator ===

Configuration:
  Requests: 100
  Window: 60s
  Burst: 10

Rates:
  1.67 requests/second
  100.00 requests/minute
  6000.00 requests/hour
  144000 requests/day

Intervals:
  Min interval between requests: 600.00ms

Token Bucket:
  Bucket size: 10
  Refill rate: 1.6667 tokens/second
  Refill interval: 600.00ms per token
  Time to refill bucket: 6.00s

Sliding Window:
  Window size: 60s
  Max in window: 100

When to use this

API clients derive pacing, gateways compare refill with burst capacity, and load tests project traffic implied by a steady quota.

Edge cases

  • A full ten-token bucket can serve ten immediate requests, so 600 ms is an average spacing rather than a mandatory delay.
  • Projecting 100 per minute to 144,000 per day assumes uninterrupted refill for 24 hours.
  • Distributed counters, clock skew, queues, tiered quotas, and server-specific HTTP 429 behavior are not modelled.

References