About Portkey
Portkey is an AI gateway and control panel for teams that run models in production. It brings routing, observability, guardrails and prompt management into one platform.
What it does
Portkey sends requests from your application to many LLM providers through a single API. A real-time dashboard shows usage and model behavior, so teams can spot anomalies early and manage usage. Guardrails check inputs and outputs, and governance tools help set access rules. Repeated requests can be cached to reduce cost. The gateway repository is MIT licensed, and the company offers managed and enterprise options on top of it.
Key features
- Route requests to 1,600+ LLMs through one API
- Real-time dashboard for usage and model behavior
- Guardrails that check inputs and outputs
- Prompt management and governance rules
- Caching of repeated requests to save cost
Who it's for
The platform suits platform teams and enterprises that run several models for many users. A small app that calls one provider may not need a gateway.
Pricing and good to know
The gateway code is open source under the MIT license, so the core can be self-hosted. The managed platform and enterprise features have their own pricing, which the site lists on a separate page. Check which features are part of the open source gateway and which belong to the paid platform before you plan a rollout.
Highlights
- Route requests to 1,600+ LLMs through one API
- Real-time dashboard for usage and model behavior
- Guardrails that check inputs and outputs
- Prompt management and governance rules
- Caching of repeated requests to save cost
