x402 Universal API Router

A framework for bridging any Web2 API to the agent economy with x402 micropayments.

// THE VISION

The agent economy needs a universal translation layer. While Web3 is built on open protocols, the vast majority of today's powerful APIs (LLMs, data sources, cloud services) are locked behind Web2 subscription models and centralized payment gateways.

Our solution is a decentralized routing protocol that allows anyone to bring any API on chain, making it instantly accessible to autonomous agents via pay per use x402 micropayments on Solana.

We are replacing rigid, expensive subscriptions with a fluid, on demand, and truly open marketplace for digital services.

// HOW IT WORKS

[ Client ]

@x402/client

AI Agent or Dev

[ Registry ]

api-registry

Discovery Layer

[ Provider ]

@x402/node

API Wrapper

[ Web2 API ]

e.g. OpenAI

Existing Service

Flow:1. Discover2. Pay & Call3. Wrap API

// VIDEO DEMO (3 MIN)

// CORE INFRASTRUCTURE: Open Source Packages

// THE DEMO: Pay Per Use Voice AI

To prove the power of our framework, we built a live, end to end voice AI agent demo. This is not a simulation; it is a real world application running on our x402 infrastructure.

[ Voice Demo Client ]

A web based client allowing real time voice conversation with an AI agent. It uses our @x402apis/client SDK to perform an x402 payment, purchasing a session token from the provider.

View Repository

[ Voice Demo Provider ]

A service running our @x402apis/node software. It wraps Deepgram and OpenAI APIs, exposing them as a single x402 enabled endpoint. It verifies payments and orchestrates the AI interaction.

View Repository
LAUNCH LIVE DEMO

You will need a Solana wallet. All transactions are x402 based.

The entire demo is open source, showcasing our tooling.

// GRADIENT PARALLAX INTEGRATION

Deepgram (STT)

Parallax (LLM)

Deepgram (TTS)

Step 1: Deepgram (STT)

User speaks. Audio is streamed to Deepgram for instant, low-latency transcription.

Achievement: Decentralized LLM is Live!

We successfully integrated Gradient's Parallax as the core LLM for our real-time voice agent. This achievement realizes our goal of building a modular, decentralized AI stack. Our architecture now leverages Deepgram for best-in-class, low-latency speech-to-text (STT) and text-to-speech (TTS), while Parallax handles the core reasoning.