GenPark LLM Token Bucket & Sliding Window Rate Limiter

A local MCP server that provides adaptive TPM/RPM rate limiting with leaky bucket queuing and burst wait estimation for LLM calls.

What you can do

• Apply multi-tier token-per-minute and request-per-minute rate limits to LLM traffic.
• Queue requests smoothly through a leaky bucket instead of rejecting them outright.
• Estimate how long a burst of requests will have to wait before it clears.
• Adapt limits across tiers as usage changes.

When to use it

Useful when an assistant or agent pipeline calls LLM APIs and needs to stay under provider TPM/RPM quotas without hard failures, or when you want predictable wait estimates for bursts of requests.

Requirements

Runs as a local stdio process on the user's machine (desktop only). The README's example configuration launches it with Python and a local script, and it is implemented entirely in Python 3 standard library, so a Python 3 runtime is needed. No environment variables, headers or authentication are declared.