Agent Rate Limiter Mcp
Share fleet-wide rate limits across A2A and many MCP servers so one agent cannot exhaust per-server quotas alone.
Overview
Agent Rate Limiter MCP is a MCP server for the Operate phase that applies fleet-wide shared rate limits across A2A and multi-MCP agent deployments.
What is this MCP server?
- Fleet-wide shared rate limiter for A2A plus multi-MCP deployments
- Addresses per-MCP-server limits that do not coordinate across agents
- Streamable HTTP remote and stdio PyPI package v1.0.4
- Python stdio transport via agent-rate-limiter-mcp on PyPI
- Remote URL: https://api.meok.ai/v1/a2a/rate-limiter
- Server version 1.0.4
- 1 streamable-http remote
- 1 PyPI stdio package (Python)
What problem does it solve?
Independent per-server rate limits let multiple agents collectively exceed safe throughput and burn API budgets without a single coordination point.
Who is it for?
Builders running several MCP tools and orchestrators who hit duplicated or uncoordinated throttles in automation-heavy workflows.
Skip if: One MCP and manual, human-paced usage where provider defaults are enough.
What do I get? / Deliverables
After install, agents share a centralized limiter so burst traffic across MCP and A2A calls is throttled before providers hard-fail your stack.
- Central MCP endpoint for fleet-wide rate decisions
- Reduced risk of uncoordinated API stampedes
- Configurable shared limits across agent traffic
Recommended MCP Servers
Journey fit
Fleet rate limiting is canonical on Operate → Infra because it stabilizes live multi-MCP deployments after individual server caps prove insufficient. Infra subphase fits shared quota coordination across processes and remotes, not feature coding.
How it compares
Shared quota MCP for agent fleets, not a single-API SDK retry wrapper.
Common Questions / FAQ
Who is agent-rate-limiter-mcp for?
Solo builders and small teams operating multi-agent, multi-MCP systems that need one place to coordinate call rates and avoid quota stampedes.
When should I use agent-rate-limiter-mcp?
Use it when two or more agents or MCP servers can hit the same external APIs or internal services concurrently, especially under autonomous loops.
How do I add agent-rate-limiter-mcp to my agent?
Add the PyPI stdio package agent-rate-limiter-mcp to your MCP configuration or use the streamable-http endpoint at api.meok.ai/v1/a2a/rate-limiter.