EU-sovereign · OpenAI-compatible AI inference

Bring your hardware.
Sell what you’re not using.

Porten routes paid AI traffic to idle machines — you earn for every token they serve, and run your own models free. One OpenAI-compatible API, with regional control when you need it.

Turn any Mac or Linux box into a node — one line:

$curl -fsSL https://porten.ai/install.sh | sh

Run it → approve in your browser → you’re earning.

Earn from idle hardware

List the capacity you’re not using. We route paid traffic to it and you earn for every token your machine serves — paid out in cash.

Run your own AI, free

Your own models run on your own hardware at zero platform fee. We handle routing, billing and uptime so you don’t have to.

One clean API

An OpenAI-compatible endpoint that drops into the SDKs you already use. Pin requests to a region — EU, US, or Asia — when data residency matters.

Run only on your own hardware, only in the EU

Regulated or public-sector? Pin inference to machines you own and the region you choose — enforced, fail-closed. We run the orchestration so it’s usable from day one.

Works with your stack

OpenAI SDKVercel AI SDKLangChainLlamaIndexLiteLLMOpenCodeContinue

Why Porten

Both sides of the network. One balance.

Rent out raw GPUs and you get no usable API. Buy from an aggregator and you can’t put your own hardware to work. Porten does both — on one balance.

Feature
Porten
GPU rental
Aggregators
OpenAI-compatible APIYesNoYes
Earn from idle capacityYesYesNo
Run your own models freeYesNoNo
Regional control (EU / US / Asia)YesPartialPartial
Paid in cash, no cryptoYesPartialNo

How it works

Your hardware, working while you’re not.

1

Install in one line

Run the installer on any Mac or Linux box. It sets itself up and connects in seconds — nothing to configure.

2

We handle the rest

Scheduling, routing, billing and uptime across the network — all managed for you.

3

Earn on idle, use it free

Spare capacity earns you money; your own inference runs free. One balance, both directions.