Skip to content

Portkey AI Gateway

A fast AI gateway with a unified API for many LLMs

Visit official site

What Portkey AI Gateway is

The gateway is described as the interface between applications and hosted LLMs. It uses a unified API to optimise requests to OpenAI, Anthropic, Mistral, LLama2, Anyscale and Google Gemini, which enables smooth routing.

What you can do with it

  • Use a unified API signature across providers
  • Route requests to multiple models
  • Retry failed requests automatically
  • Balance load and fall back on errors

Example uses

  • Call different LLMs through the gateway with the OpenAI SDK
  • Configure a fallback strategy to prevent request failure
  • Set multiple API keys for an application to balance load

Application scenarios

  • Use various LLMs through a unified API
  • Improve the stability of an application
  • Optimise how different LLMs are used

Pros & cons

✓ What we like

  • Built-in retries support application reliability
  • Multi-model load balancing is included

! What to watch out for

  • Requests are routed through the gateway rather than sent directly to providers
  • The named reliability behaviour relies on built-in retries and fallback

FAQ

Which providers are named?

The material names OpenAI, Anthropic, Mistral, LLama2, Anyscale and Google Gemini.

What reliability features are listed?

Automatic retries, load balancing and fault-tolerant fallback are listed.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review