Budget Concurrency Against Provider Rate Limits

Use when sending many model calls in parallel. Size concurrency to the tier's RPM/TPM limit, back off on 429, and bound the queue so a batch does not stall or overrun cost.

jiayaoqijia Updated 81 repo stars

File contents

jiayaoqijia/cryptoskill commit 1dd1e3ca75

Frequently asked questions

npx get-skill add jiayaoqijia/budget-concurrency-against-provider-rate-limits