Save the lookup beside the citation

A historical page check has two inputs worth keeping: the public URL and the date you meant to inspect. Save both with the citation record. A later run can ask for the same archived point instead of guessing which capture mattered.

archive-snapshot returns the closest capture URL and its archive timestamp. Set fetch_text when the next step needs cleaned page text for a quote check. 

curl -X POST https://x402.agentutility.ai/archive-snapshot \
  -H "Content-Type: application/json" \
  -d '{"url":"https://www.example.com/policy","timestamp":"2024-06-01","fetch_text":true}'

The first response requests x402 payment. Keep the request and returned JSON together, then reuse the same body for the next citation review. 

{
  "citation_id": "policy-source-2024-06",
  "request": {
    "url": "https://www.example.com/policy",
    "timestamp": "2024-06-01",
    "fetch_text": true
  },
  "snapshot_url": "https://web.archive.org/web/...",
  "snapshot_iso_timestamp": "2024-06-01T00:00:00Z"
}

Use a date when the claim depends on a past version of a page. Leave it out when the workflow needs the most recent available capture.

Read the result before starting a new lookup

A 404 means both archive checks answered without a capture. A 424 means the archive could not give a dependable answer yet. Keep the same saved request for a later retry in that case, since a different date changes the research question.

The 2026-11-02 recovery snapshot found two archive-snapshot buyers, including one repeat buyer that had gone quiet. They made two calls worth 0.04 USDC. This page keeps the next paid pass tied to the original URL and date.