Decoder. plain-English AI glossary

Rate limiting

● Core

Also called Rate Limiting

Capping how many requests someone can make in a window, to blunt abuse, runaway costs, and attacks.

Think of it like

A turnstile that only lets so many people through per minute, so no one can rush the gate.

Example

An API returns "429 Too Many Requests" after a script tries 10,000 generations in a minute, throttling a scraping or brute-force attempt.

How it actually works

Limits are set per key, user, or IP and enforced with algorithms like token bucket or sliding window. Beyond fairness and cost control, rate limiting slows brute-force jailbreak search and automated abuse by making each attempt cost time. It is a blunt but reliable safety and reliability primitive.

For product teams

A simple lever that protects cost, reliability, and safety all at once.

For engineers

Enforce per-identity quotas via token-bucket/sliding-window; it also raises the cost of automated jailbreak and abuse campaigns.

Related

Read anything AI without the jargon

Look up any term in plain English, or save terms as you read with the free Chrome extension.

Open DecoderAdd to Chrome