GitHub Repository

Fastest enterprise AI gateway (50x faster than LiteLLM) with adaptive load balancer, cluster mode, guardrails, 1000+ models support & <100 µs overhead at 5k RPS.

6,220 starsGo

Fastest Enterprise AI Gateway

Name: Fastest Enterprise AI Gateway
Availability: InStock
Author: aanthonymax

by aanthonymax·Jul 2, 2026·2 points·0 comments

Visit Project View on HN

Similar Projects

Infrastructure●●Solid

Bifrost: Fastest enterprise AI gateway

Bifrost combines an OpenAI-compatible front door with adaptive load balancing, semantic caching, automatic failover, cluster mode and a built-in web UI — you can spin it up with npx or Docker in seconds. The performance claims (sub-100µs overhead at 5k RPS, '50x faster than LiteLLM') and multi-provider routing are the project's selling points; I want to see independent benchmarks and deeper docs on guardrails/provider quirks before trusting it for critical workloads.

WizardrySolve My ProblemSlick

aanthonymax

104mo ago

Infrastructure●●Solid