Portkey

AI gateway for routing, securing, and observing LLM and agent requests.

Overview

Portkey is an AI gateway and observability platform for controlling requests to language models and agent systems. It centralizes provider routing, fallbacks, retries, access controls, and telemetry. Its Agent Gateway adds an agent registry, MCP server management, and role-based controls, while observability tracks requests, costs, latency, and errors.

Key features

Multi-provider AI gateway
Routing and fallback controls
Request retries and load balancing
Agent registry
MCP server management
Role-based access controls
LLM and agent observability

Use cases

Show all 10Show fewer

Alternatives

Azure AI FoundryMicrosoft Azure platform to build, deploy, and govern AI apps and agents with models and tools.AI toolPaid
IBM watsonx.aiIBM studio to build and run predictive, optimization, and generative AI across hybrid infra.AI toolFree trial
LangSmithPlatform for observing, evaluating, and deploying ai agents and LLM applications.ai agent framework
Model Context Protocol (MCP)Open protocol for connecting AI applications with tools, data sources, and prompts.ai agent framework
NVIDIA NeMoAI agent platform and toolkit for building, deploying, and managing agent workflows.ai agent framework
PezzoOpen-source LLMOps platform for managing prompts, versions, and application observability.ai agent framework

Last updated 2026-09-16