AgentPay x402 — Crypto Data Tools for AI Agents
AgentPay — x402 crypto data gateway on Stellar. 10 live pay-per-call tools: token prices, whale activity, gas tracker, DeFi TVL, Fear & Greed, Dune queries, token security. Agents pay USDC on Stellar. No API keys. Budget-aware sessions.
README
AgentPay — your agent is only as smart as its data.
AgentPay is an open x402 payment gateway that lets AI agents autonomously access real crypto data tools using USDC on Stellar or Base.
No subscriptions. No API keys. No human in the loop. Agents discover tools, pay per call ($0.001–$0.005), and get real data back — all within a hard budget cap.
→ 12 live tools: token prices, whale activity, gas tracker, DeFi TVL, Fear & Greed, yield scanner, funding rates, token security, Dune queries and more → Budget-aware Session: agents estimate costs, track spend, never exceed budget → x402 protocol: works with any x402-compatible agent → Two payment networks: Stellar (5s, $0.00001 fee) or Base mainnet USDC (2s, $0.0001 fee)
Try it in 60 seconds:
curl https://gateway-production-2cc2.up.railway.app/faucet
Or open the browser faucet: https://gateway-production-2cc2.up.railway.app/faucet/ui
Live gateway: https://gateway-production-2cc2.up.railway.app
Quickstart — 3 steps
Step 1: Get a funded test wallet
One call gives you a ready-to-use Stellar testnet wallet with 5 USDC pre-loaded:
curl https://gateway-production-2cc2.up.railway.app/faucet
Or use the browser faucet — click "Get Test Wallet", copy the snippet, run it.
Step 2: Create a Session with a budget
from agent.wallet import AgentWallet, Session, BudgetExceeded
wallet = AgentWallet(
secret_key="S...", # your Stellar secret key
network="testnet", # or "mainnet"
)
GATEWAY = "https://gateway-production-2cc2.up.railway.app"
with Session(wallet=wallet, gateway_url=GATEWAY, max_spend="0.05") as session:
print(f"Balance: {wallet.get_usdc_balance()} USDC")
print(f"Budget: {session.remaining()} remaining")
The Session enforces a hard USDC cap across all calls. It raises BudgetExceeded before any payment goes out if a tool would push you over, and automatically falls back to the next-cheapest tool in the same category when the preferred one is too expensive.
Step 3: Call tools — payment is automatic
with Session(wallet=wallet, gateway_url=GATEWAY, max_spend="0.05") as session:
# Token price — $0.001
r = session.call("token_price", {"symbol": "ETH"})
print(f"ETH: ${r['price_usd']:,.2f} ({r['change_24h_pct']:+.2f}% 24h)")
# Fear & Greed Index — $0.001
r = session.call("fear_greed_index", {"limit": 1})
print(f"Sentiment: {r['value']}/100 — {r['value_classification']}")
# DeFi TVL — $0.002
r = session.call("defi_tvl", {"protocol": "aave"})
print(f"Aave TVL: ${r['tvl'] / 1e9:.1f}B ({r['change_1d']:+.1f}% 24h)")
# Crypto news — $0.003
r = session.call("crypto_news", {"currencies": "ETH", "filter": "hot"})
for h in r["headlines"][:3]:
print(f" [{h['sentiment']:>7}] {h['title'][:55]}")
print(f"\nTotal spent: {session.spent()}")
print(f"Remaining: {session.remaining()}")
Each session.call() handles the full x402 flow internally:
- Checks your remaining budget against the tool's price (pre-flight, no payment yet)
- POSTs to the gateway, receives a
402with{payment_id, amount_usdc, pay_to} - Sends USDC on Stellar — ~2–3 seconds on testnet
- Retries the request with
X-Payment: tx_hash=<hash>,from=<addr>,id=<payment_id> - Returns the data
Available Tools
| Tool | Price | Parameters | Returns |
|---|---|---|---|
token_price |
$0.001 | symbol (BTC, ETH, SOL…) |
price_usd, change_24h_pct, market_cap_usd |
gas_tracker |
$0.001 | — | slow/standard/fast gwei, base_fee_gwei |
fear_greed_index |
$0.001 | limit (days of history, default 1) |
value 0–100, value_classification, history[ ] |
wallet_balance |
$0.002 | address, chain (ethereum/stellar) |
token balances |
whale_activity |
$0.002 | token, min_usd (default 100k) |
large_transfers[ ], total_volume_usd |
defi_tvl |
$0.002 | protocol (optional, e.g. "uniswap") |
tvl, change_1d, change_7d, chains[ ] |
token_security |
$0.002 | contract_address, chain (ethereum/bsc) |
risk_level, is_honeypot, buy_tax, sell_tax, holder_count |
yield_scanner |
$0.004 | token, chain (optional), min_tvl (default $1M) |
top 10 pools by APY with protocol, tvl_usd, risk_level |
dex_liquidity |
$0.003 | token_a, token_b |
volume_24h_usd, market_cap_usd, ath_usd |
funding_rates |
$0.003 | asset (optional, e.g. "BTC") |
funding_rate_pct, annualized_rate_pct, sentiment per exchange |
crypto_news |
$0.003 | currencies (e.g. "ETH,BTC"), filter (hot/new/rising) |
headlines[ ] with title, url, sentiment, score |
dune_query |
$0.005 | query_id, limit (default 25) |
rows[ ], columns[ ], row_count from Dune Analytics |
Discover all tools dynamically:
import httpx
tools = httpx.get(f"{GATEWAY}/tools").json()["tools"]
for t in tools:
print(f"{t['name']:<22} ${t['price_usdc']} — {t['description']}")
Payment Options
AgentPay accepts USDC payments on two networks:
- Stellar — $0.00001 per tx, 5-second settlement (recommended for agents using the Python SDK)
- Base — $0.0001 per tx, 2-second settlement (EIP-3009
transferWithAuthorizationon Base mainnet)
The gateway's 402 response advertises both options simultaneously. Clients pick the network that suits them — no configuration required on the tool side.
{
"payment_options": {
"stellar": { "pay_to": "G...", "amount_usdc": "0.001", ... },
"base": { "network": "eip155:8453", "asset": "0x833589f...", "amount_usdc": "0.001", ... }
}
}
Base payments use Mode B direct on-chain settlement: the client calls transferWithAuthorization on the USDC contract and sends the tx_hash in the PAYMENT-SIGNATURE header. The gateway verifies the receipt via JSON-RPC.
MCP Server
AgentPay ships a Model Context Protocol server that gives Claude Desktop direct access to all 12 tools. Payments happen automatically in the background using Stellar USDC.
See README_MCP.md for setup instructions.
Quick config (claude_desktop_config.json):
{
"mcpServers": {
"agentpay": {
"command": "python",
"args": ["/path/to/agentpay/gateway/mcp_server.py"],
"env": {
"STELLAR_SECRET_KEY": "S...",
"AGENTPAY_GATEWAY_URL": "https://gateway-production-2cc2.up.railway.app"
}
}
}
}
Without the SDK — raw HTTP
The x402 flow works with any HTTP client in any language.
GATEWAY="https://gateway-production-2cc2.up.railway.app"
AGENT_ADDR="G..." # your Stellar public key
# 1. Call the tool → receive 402 payment challenge
RESPONSE=$(curl -s -X POST "$GATEWAY/tools/token_price/call" \
-H "Content-Type: application/json" \
-d "{\"parameters\":{\"symbol\":\"ETH\"},\"agent_address\":\"$AGENT_ADDR\"}")
PAYMENT_ID=$(echo $RESPONSE | python3 -c "import sys,json; print(json.load(sys.stdin)['payment_id'])")
AMOUNT=$(echo $RESPONSE | python3 -c "import sys,json; print(json.load(sys.stdin)['amount_usdc'])")
PAY_TO=$(echo $RESPONSE | python3 -c "import sys,json; print(json.load(sys.stdin)['pay_to'])")
# 2. Send USDC on Stellar (memo = first 28 chars of payment_id)
TX_HASH=$(python3 -c "
from stellar_sdk import Keypair, Server, Network, Asset, TransactionBuilder
kp = Keypair.from_secret('S...')
server = Server('https://horizon-testnet.stellar.org')
acct = server.load_account(kp.public_key)
tx = (TransactionBuilder(acct, Network.TESTNET_NETWORK_PASSPHRASE, base_fee=100)
.add_text_memo('$PAYMENT_ID'[:28])
.append_payment_op('$PAY_TO', Asset('USDC','GBBD47IF6LWK7P7MDEVSCWR7DPUWV3NY3DTQEVFL4NAT4AQH3ZLLFLA5'), '$AMOUNT')
.set_timeout(30).build())
tx.sign(kp)
print(server.submit_transaction(tx)['hash'])
")
# 3. Retry with payment proof → get data
curl -s -X POST "$GATEWAY/tools/token_price/call" \
-H "Content-Type: application/json" \
-H "X-Payment: tx_hash=$TX_HASH,from=$AGENT_ADDR,id=$PAYMENT_ID" \
-d "{\"parameters\":{\"symbol\":\"ETH\"},\"agent_address\":\"$AGENT_ADDR\"}"
Session API reference
# Pre-check price without paying
price = session.estimate("dune_query") # "$0.005"
# Check headroom before committing
if not session.would_exceed("0.005"):
result = session.call("dune_query", {"query_id": 3810512, "limit": 10})
# Access spend state mid-session
session.spent() # "$0.004"
session.remaining() # "$0.046"
# Full breakdown after session closes
summary = session.summary()
# {
# "calls": 3,
# "spent_usdc": "0.006",
# "spent_fmt": "$0.006",
# "remaining_fmt": "$0.044",
# "breakdown": [
# {"tool": "token_price", "amount_usdc": "0.001", "tx_hash": "abc123..."},
# {"tool": "gas_tracker", "amount_usdc": "0.001", "tx_hash": "def456..."},
# {"tool": "defi_tvl", "amount_usdc": "0.002", "tx_hash": "ghi789..."}
# ]
# }
Run the demo
git clone <this-repo> && cd agentpay
python -m venv venv && source venv/bin/activate
pip install -r requirements.txt
# Copy env and add your Stellar secret key
cp .env.example .env
# Run the full 5-tool ETH analysis against the live gateway
AGENTPAY_GATEWAY_URL=https://gateway-production-2cc2.up.railway.app \
python agent/budget_demo.py
Expected output: ETH price, gas, DEX liquidity, whale moves, and Dune onchain data — all paid for autonomously in ~$0.012 USDC.
Architecture
agent (Python SDK)
│
│ POST /tools/{name}/call
│ ← 402 {payment_options: {stellar: {...}, base: {...}}}
│ → Stellar USDC payment (~3s) OR Base transferWithAuthorization (~2s)
│ → retry with X-Payment or PAYMENT-SIGNATURE header
│ ← 200 {result: ...}
▼
gateway (FastAPI on Railway)
│
├── registry/registry.py — 12-tool catalog with prices & dev wallets
├── gateway/stellar.py — Stellar payment verification via Horizon
├── gateway/base.py — Base payment verification via JSON-RPC
└── gateway/main.py — real API dispatchers
├── CoinGecko token_price, dex_liquidity
├── Etherscan V2 gas_tracker, whale_activity, wallet_balance
├── DeFiLlama defi_tvl
├── alternative.me fear_greed_index
├── Reddit crypto_news
├── Dune Analytics dune_query
├── GoPlus token_security
├── DeFiLlama yield_scanner
└── Binance/Bybit/OKX funding_rates
Fee model: Gateway charges 15% (GATEWAY_FEE_PERCENT=0.15), forwards the rest to each tool developer's Stellar wallet. All payments settle on-chain in ~2–5 seconds.
Note: AgentPay currently uses the x402 pay-first pattern with classic Stellar PAYMENT ops. OZ Facilitator (verify-first, Soroban SAC) migration planned for v2.
Agent Discovery
AgentPay is discoverable by autonomous agents at standard discovery paths:
- AgentPay manifest:
https://gateway-production-2cc2.up.railway.app/.well-known/agentpay.json - A2A agent card:
https://gateway-production-2cc2.up.railway.app/.well-known/agent.json
Any x402-compatible agent can discover and use AgentPay tools without human setup.
All 12 AgentPay tools are also indexed on x402scout under network: stellar-testnet.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.