Portkey

LLM observability and an AI gateway — logging, routing, and guardrails.

EstablishedMed lock-in
Visit Website ↗Compare ⇄

Pricing

Free tier

Usage-based

Adoption

→Stable

License

Proprietary

Data freshness

Unverified

Overview

What is Portkey?

Portkey is an AI gateway and observability layer for LLM apps: request logging/metrics, smart routing/fallbacks across providers, caching, and guardrails.

Key differentiator

“An AI gateway that pairs LLM observability with routing and guardrails.”

Capability profile

Capability Radar

Ease of StartEcosystemValueMaturityFlexibilityScale Ready

Honest assessment

Strengths & Weaknesses

↑ Strengths

AI gateway + observabilityhigh

One layer for logging, routing, fallbacks, and caching

Multi-provider routing/fallbacksmedium

Resilience across LLM providers

Guardrails and cachingmedium

Control cost and safety

↓ Weaknesses

Adds a proxy hoplow

Gateway sits in the request path

Younger ecosystemlow

Newer than general APM

Fit analysis

Who is it for?

✓ Best for

Monitoring LLM API calls and costs

Logging, metrics, and routing for LLMs

Multi-provider LLM resilience

Fallbacks and caching

Cost structure

Pricing

Free Tier

Available

Free tier with monthly request allowance

Starts at

Usage-based

Model

Usage-based

Enterprise

Available

Performance benchmarks

How Fast Is It?

Ecosystem

Relationships

Alternatives

HeliconeLangSmith

Next step

Get Started with Portkey

Step-by-step setup guide with code examples and common gotchas.

View Setup Guide →