Rate limiting

#How rate limiting works

The server counts requests for each client, key or address over a time window. Requests within the limit succeed. Requests over the limit get a 429 response. The client then has to slow down and wait before it tries again.

#Why it matters for AI agents

An agent can send many requests in a second when it follows a multi-step task. Pacing requests avoids blocks and avoids putting load on the service. A polite client treats a 429 as a signal to slow down, not as a failure to work around.

#Wield and rate limiting

Wield limits every run: 12 steps, 40 requests, 5 MB per response and 60 seconds per run. The extension also paces requests per site across all clients. Time spent waiting for your Allow on a write does not count toward the 60 seconds.

#Related terms

• Web scraping
• Akamai Bot Manager
• Site challenge
• Internal API

#FAQ

#What does HTTP 429 mean?

429 Too Many Requests means the client sent too many requests in a given time, and the server is refusing more for now. The response can include a Retry-After header that says how long to wait.

#How should an agent respond to 429?

Stop sending requests to that site, wait for the time in Retry-After when it is given, and then retry at a slower pace. Do not try to route around the limit with more clients or other tricks.