endpointrecord
Token safety verdicts for autonomous agents.
One POST. Two cents. About five seconds.
An agent asks before it buys.
Ask
Your agent sends the token address to POST /scan. No signup, no key.
Pay
The endpoint answers 402 with an x402 offer. The agent signs 0.02 USDC on Base. The facilitator settles, gas included.
Know
SAFE, CAUTION or SCAM, with the evidence behind it and a settlement receipt.
Seven agents. One verdict.
Mint functions, hidden owners, proxies, self-destruct, blacklist switches, taxes.
Static code analysis and a simulated buy and sell. When the two disagree, we say so.
DEX pools, depth and volume. Can the token actually be sold?
Established listings, rank, market cap and age.
Known scam wallets, deployers and phishing domains.
Website and project signals matched against known fraud patterns.
Independent coverage, or the lack of it.
An incomplete check is flagged with checkIncomplete. Missing data never reads as clean.
One endpoint.
# 1. ask curl -i -X POST https://api.endpointrecord.com/scan \ -H "content-type: application/json" \ -d '{"requestKey":"<random-secret>", "tokenAddress":"0x532f…42e4"}' HTTP/2 402 payment-required: eyJ4NDAyVmVyc2lvbiI6Mi… # 2. sign the offer, repeat with PAYMENT-SIGNATURE: <base64 x402 payload> HTTP/2 200 payment-response: eyJzdWNjZXNzIjp0cnVl…
{
"verdict": "CAUTION",
"score": 6,
"contract": {
"ownerRenounced": true,
"hasMint": false,
"liquidityLocked": false,
"checkIncomplete": false,
"honeypotVerdict": {
"honeypot": false,
"sources": 2,
"consensus": true
}
},
"flags": [
"Liquidity not locked — developer
can pull liquidity (rug pull)"
]
}
Any x402 v2 client works. Full request and response schema in openapi.json.
- per verdict, paid on Base via x402
- no account, no subscription, no API key
- same
requestKey+ address within 24 h: never charged twice - settlement in progress?
503, retry free with the same key