Portkey

AI gateway with caching, fallbacks, and observability.

Visit site GitHub
Open source Self-hostable freemium
LLM gatewayPlatform engineer
LicenseMIT
PriceFree tier
UsageNo model cost
Model supportmulti
Languagestypescript
GitHub stars12.3k
Last activity2026-05-25
Verified2026-06-28
Verified byseed

An open-source AI gateway that sits between your app and LLM providers, adding response caching, automatic fallbacks, request routing, rate limiting, and unified observability across 1600+ models.

Strengths

  • Caching and fallbacks that survive provider outages
  • Unified logging across all providers in one dashboard
  • Self-hostable gateway with a managed cloud option

Tradeoffs

  • Newer ecosystem than LiteLLM
  • Managed cloud is the revenue layer; OSS is the gateway core

Covers

Self-hosting