🔍Page Claim Verifier
Verify a claim about a web page's content: fetches the page (SSRF-guarded) and reports whether the expected text appears, with the surrounding excerpt as evidence. The deterministic way for an agent to confirm 'the site says X' before acting on it.
🔌 Server-side tool
This tool runs on the server (it fetches third-party URLs directly and cannot run in your browser). Use the paid JSON API below — the worked examples show real server-side results and refresh automatically.
Worked examples
Does example.com mention documentation?
GET /api/v1/dev/page-verify?url=https%3A%2F%2Fexample.com&expect=documentation
Result: ✓ SUPPORTED — "documentation" found on the page
Inline HTML check
GET /api/v1/dev/page-verify?html=%3Chtml%3E%3Cbody%3E30-day%20money-back%20guarantee%20on%20all%20plans%3C%2Fbody%3E%3C%2Fhtml%3E&expect=money-back
Result: ✓ SUPPORTED — "money-back" found on the page
Machine API (x402)
$0.002 / callThis tool is also a JSON API for AI agents. Requests without payment receive 402 Payment Required plus instructions; agents pay USDC on Base via the x402 protocol — no accounts, no API keys.
GET /api/v1/dev/page-verify?url=https%3A%2F%2Fexample.com&expect=documentation HTTP/1.1
Host: agenttools-hub.vercel.app
→ 402 (payment required, instructions in headers)
→ 200 (after X-PAYMENT header; JSON body below)
{
"tool": "dev/page-verify",
"input": {"url":"https://example.com","expect":"documentation"},
"result": { "value": 1, "answer": "✓ SUPPORTED — "documentation" found on the page" }
}Agent docs: /llms.txt · OpenAPI spec · integration guide
About this tool
Before an agent relies on 'the website says X', this fetches the page and checks — returning SUPPORTED or NOT FOUND with the surrounding excerpt as evidence. Deterministic and auditable, unlike asking a model whether it remembers the site saying something.
Frequently asked questions
What kinds of claims can this verify?
Claims about page CONTENT: pricing mentions, policy phrases, feature names, guarantees. It cannot verify truth about the physical world — only what the page literally contains.
JavaScript-rendered pages?
No — like all server-side fetchers it reads served HTML. Most pricing/policy claims live in static HTML or metadata.