Command Palette

Search for a command to run...

UnylyUnyly
Весь каталог

Quanta X402

БесплатноНе проверен

Pay-per-call structured data for autonomous AI agents. x402-metered, MCP-native.

GitHubEmbed

Описание

Pay-per-call structured data for autonomous AI agents. x402-metered, MCP-native.

README

Pay-per-call structured data for autonomous AI agents. x402-metered, MCP-native.

A small, real reference service: a FastAPI API serves a structured dataset, every data call is metered by the x402 protocol (a USDC micropayment on Base or Solana), and the same data is exposed as MCP tools so an agent can discover and call it natively. Backed by Postgres/Supabase (SQLite locally).

This is a working slice of the agent-commerce stack: agent -> MCP tool -> x402 paywall -> data.


Why it exists

It's the exact shape of the systems I build for: a live API that sells structured data to autonomous agents, monetized per call, with an MCP front door. It runs the real protocols (official x402 SDK + official MCP SDK), not a mock.


Architecture

                      ┌──────────────────────────────────────────┐
   AI agent  ───────▶ │  MCP tools (FastMCP)   /mcp              │
   (Claude, etc.)     │    list_assets · get_asset · get_signal  │
                      └───────────────┬──────────────────────────┘
                                      │  same data layer
   HTTP client ─────▶ ┌──────────────▼──────────────────────────┐
   (curl, SDK)        │  FastAPI  /v1/*   ── x402 middleware ────┤
                      │    402 challenge ▶ verify ▶ settle ▶ data │
                      └───────────────┬──────────────────────────┘
                                      │
                      ┌───────────────▼──────────────┐   ┌──────────────────┐
                      │  data.py (single source)      │   │  x402 facilitator │
                      │  assets + usage_log           │   │  verify + settle  │
                      └───────────────┬──────────────┘   │  Base / Solana    │
                                      │                   └──────────────────┘
                      ┌───────────────▼──────────────┐
                      │  Postgres / Supabase (SQLite) │
                      └───────────────────────────────┘

The x402 request loop

  1. Agent calls a /v1/* route with no payment.
  2. Server replies 402 Payment Required + the accepted rails (USDC on Base Sepolia and Solana devnet), price, and pay-to address.
  3. Agent signs a payment and retries with an X-PAYMENT proof.
  4. The facilitator verifies + settles on-chain; the server returns the data and writes a row to usage_log (who paid, which chain, which record).

Metering is a feature flag. If the x402 wheel isn't present (local dev on an unsupported Python), the app logs a warning and serves unmetered so you can still work on everything else. It never fails to boot.

Layout

app/
  main.py        FastAPI app, /v1/* data routes, x402 wired in
  x402_setup.py  multi-chain (EVM + Solana) metering, import-safe
  data.py        seed data + shared queries + usage log (one source of truth)
  models.py      Asset + UsageLog (SQLAlchemy 2.0 async)
  db.py          async engine; DATABASE_URL swaps SQLite <-> Supabase
  config.py      typed settings
mcp_server/
  server.py      FastMCP tools (x402-gated per tool), stdio or HTTP
client/
  pay_and_fetch.py  agent demo: unpaid -> 402, then paid -> data
scripts/seed.py  create tables + load sample data
tests/test_api.py
Dockerfile · railway.json

Run it locally

python3.12 -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env            # defaults to SQLite + testnet
python -m scripts.seed
uvicorn app.main:app --port 4021

Prove the metering is live:

curl -i http://localhost:4021/v1/assets/BTC      # -> 402 Payment Required + rails
curl    http://localhost:4021/internal/usage     # -> audit trail (free)
python -m client.pay_and_fetch                   # unpaid 402, then paid loop

MCP (stdio, for Claude Desktop / MCP Inspector):

python -m mcp_server.server

Deploy

  • Railway / Cloud Run / Render: Docker build (Python 3.12). railway.json is included; health check is /health.
  • Database: set DATABASE_URL to your Supabase Postgres string (postgresql+asyncpg://...).
  • Payments: testnet works with the public x402.org facilitator and no keys. For mainnet settlement, set CDP_API_KEY_ID / CDP_API_KEY_SECRET (Coinbase CDP) and switch the network.

Honest scope

  • The dataset is a clearly-labelled sample snapshot, not a live market feed - enough to prove the pay-per-call mechanics end to end.
  • Settlement runs on testnet (Base Sepolia / Solana devnet). Flipping to mainnet is config, not code.
  • The full paid loop in client/pay_and_fetch.py needs a funded test wallet; the unpaid 402 proves the contract without one.

Stack

FastAPI · SQLAlchemy 2.0 (async) · Postgres/Supabase · official x402 SDK (EVM + SVM) · official MCP SDK (FastMCP) · Docker.

from github.com/themeknock/quanta-x402-mcp

Установка Quanta X402

У этого сервера нет опубликованного пакета — он собирается из исходников. Открой репозиторий и следуй инструкции в README.

▸ github.com/themeknock/quanta-x402-mcp

FAQ

Quanta X402 MCP бесплатный?

Да, Quanta X402 MCP бесплатный — установка в пару кликов через Unyly без оплаты.

Нужен ли API-ключ для Quanta X402?

Нет, Quanta X402 работает без API-ключей и переменных окружения.

Quanta X402 — hosted или self-hosted?

Self-hosted: сервер запускается локально на твоей машине командой из раздела установки.

Как установить Quanta X402 в Claude Desktop, Claude Code или Cursor?

Открой Quanta X402 на unyly.org, выбери вкладку своего клиента (Claude Desktop, Claude Code, Cursor) и нажми Install — конфиг сгенерируется автоматически, без правки JSON.

Похожие MCP

Compare Quanta X402 with

Не уверен что выбрать?

Найди свой стек за 60 секунд

Автор?

Embed-бейдж для README

Похожее

Все в категории ai