A peer-to-peer network
that routes AI inference
without a middleman.
Loading...

Peer-to-peer
INFERENCE
NO MIDDLEMAN

Users and model providers connect directly. Run a private router on your machine, point your tools to localhost and pay providers on-chain. It’s the open market AI access should have been.

Understanding
What is
Elyra?

Virtual private router

A self-hosted router on your machine with a localhost endpoint. Any OpenAI-compatible tool works unchanged.

Peer-to-peer routing

Built to reach 600+ models from 150+ providers (launch target), chosen by price, speed, reputation or privacy.

Direct on-chain payment

Usage settles straight to the provider. No platform balance, no markup.

How it Works

Routing
Your rules

Install the router, set a policy and let it find the best peer for every request. Pay the provider directly.

01
Connecting

Point your tools

Set the base URL to localhost in any OpenAI-compatible tool. No code changes, no new SDK, no platform account.

02
Routing

Set your policy

Weight price, speed, reputation and privacy. The VPR picks the best peer for every single request.

03
Settling

Pay peers directly

Usage settles to the provider via an escrow contract and signed vouchers on Robinhood Chain (EVM L2, ETH gas). TEE peers are optional and self-declared.

Your advantage

Why builders
choose Elyra

Direct settlement

Peers are paid directly, with no platform balance or markup. Settlement targets Robinhood Chain (EVM L2, ETH gas) via an escrow contract and signed vouchers.

Elyra
Elyra
Elyra
Elyra
Elyra
Elyra

Join the
network

Elyra is built for the people who use AI and the people who supply it.

Get Started
Read the docs
Drag
ai
01

Developers

Point any OpenAI-compatible tool at localhost and route by cost or speed, with no code changes.

ai
02

Providers

Own a GPU or API access? Earn per token, settle on-chain and join without approvals.

ai
03

Privacy
teams

Route only to attested TEE sellers you verified, or keep it local, so prompts skip the platform.

ai
04

Agents

Autonomous software buys inference per request from the cheapest healthy peer.

ai
01

Developers

Point any OpenAI-compatible tool at localhost and route by cost or speed, with no code changes.

ai
02

Providers

Own a GPU or API access? Earn per token, settle on-chain and join without approvals.

ai
03

Privacy
teams

Route only to attested TEE sellers you verified, or keep it local, so prompts skip the platform.

ai
04

Agents

Autonomous software buys inference per request from the cheapest healthy peer.

ELYRA

Inference should be a market, not a platform.

Get started
Run a node
Read the docs
© 2026 Elyra. All Rights Reserved.