Claude API usage limits, rate limits and how to raise them

What's happening

A 429 response means you crossed a per-minute limit — requests, input tokens or output tokens — not that your account is suspended.

Fix it

1. Read the rate limit headers on the response to see which limit was hit and when it resets.
2. Queue and pace requests instead of bursting them in parallel.
3. Trim system prompts and retrieved context to reduce input tokens per minute.
4. Move to a higher usage tier by purchasing prepaid credits.
5. Request a limit increase through the Console for sustained production volume.

Still stuck?

Batch processing is a better fit than raising limits when the work is not latency sensitive.