Safe web access for AI agents.
Read any page as clean Markdown, search the live web, get grounded answers, fact-check claims and check a token before you trade it — with prompt-injection screening on every page and every quote checked verbatim against its source. Pay per call; you're only charged when it works.
Endpoints
Safe Read $0.002
Fetch one public web page and return clean, LLM-ready Markdown with title, word count, content hash and a prompt-injection risk scan of the page text. Use it when an agent needs to read a URL without ingesting scripts, boilerplate or hidden instructions.
Web Search $0.004
Live web search for agents. Returns ranked results with title, URL and snippet; optionally reads the top results into clean Markdown in the same call. Use it to find current sources before answering or citing.
Grounded Answer $0.015
Answer a question from live web sources. AxonGate searches, reads the best pages, and returns a concise answer with numbered citations whose quotes are checked verbatim against the fetched pages. Use it when an agent needs a sourced answer, not just links.
Claim Verify $0.02
Fact-check one claim against live web sources (or URLs you supply). Returns a verdict (supported, refuted, mixed, insufficient), a confidence score and supporting or contradicting quotes that are checked verbatim against the fetched pages. Use it before an agent repeats a claim.
Injection Scan $0.002
Scan untrusted text (a web page, email, tool output or document) for prompt-injection and agent-hijacking patterns before it reaches an LLM. Returns a risk level, score, findings with offsets and a sanitized copy. Deterministic, sub-50ms.
Token Check $0.01
Safety verdict for a crypto token before an agent buys, swaps or accepts it: honeypot, sell and buy taxes, owner powers (mint, blacklist, balance changes), holder concentration and DEX liquidity, combined into avoid / caution / ok with reasons. EVM chains and Solana.
Batch Read $0.03
Read up to 20 public URLs in one paid call. Each page comes back as clean Markdown with its own prompt-injection scan; pages that fail are reported per URL without failing the batch. Use it to load a reading list or search results in one round trip.
Try it in one call
# Any x402 client works. With Node:
npm i @x402/fetch @x402/evm viem
import { wrapFetchWithPayment, x402Client } from "@x402/fetch";
import { registerExactEvmScheme } from "@x402/evm/exact/client";
import { privateKeyToAccount } from "viem/accounts";
const client = new x402Client();
registerExactEvmScheme(client, { signer: privateKeyToAccount(process.env.PRIVATE_KEY) });
const pay = wrapFetchWithPayment(fetch, client);
const res = await pay("https://api.axongate.one/v1/verify", {
method: "POST", headers: { "content-type": "application/json" },
body: JSON.stringify({ claim: "HTTP status code 402 is reserved for future use." }),
});
console.log(await res.json()); // verdict, confidence, verbatim-checked evidenceWithout payment, every endpoint answers 402 Payment Required with machine-readable requirements in the PAYMENT-REQUIRED header.
Why agents use AxonGate
Injection-screened reads
Every page is scanned for instruction overrides, role hijacks, secret-exfiltration lures, hidden Unicode and data-leaking image links before your model sees it.
Quotes that actually exist
Answers and verdicts cite quotes that are matched verbatim against the fetched page. Anything the model can't back up is dropped, and verdicts without proof are downgraded.
Pay only for success
Settlement happens after the response. Timeouts, blocked targets and supplier errors return 4xx/5xx and are never charged.
No account, no key
Discover, pay and call in one request cycle. Works with any x402 v2 client and appears in x402 directories.