Agent Link Checker

Check HTTP status and redirect evidence before a URL enters an automated workflow.

Send up to 100 public URLs. Receive one structured result per unique URL, including the final address, HTTP status, redirect chain, check time and an explicit outcome category. Useful for checking links in generated reports, reviewing source lists and validating URLs before a larger crawl.

What it checks

• Responding HTTP endpoints and redirects.
• Missing pages: HTTP 404 and 410.
• Authentication, denied access and rate limits as separate outcomes.
• Temporary server failures, DNS/TLS/network failures and redirect loops.
• Duplicate URLs, including fragment-only variants, once per run.

A responding endpoint is not proof of the content or truth of a citation. This tool does not inspect page text, validate quotations, identify soft 404s, evaluate reputation or judge source accuracy.

Input

{
  "urls": ["https://example.com/", "https://www.iana.org/domains/reserved"],
  "timeoutSeconds": 8,
  "maxRedirects": 5
}

The maximum is 100 input URLs. Each URL has a total time limit of 2–15 seconds, including DNS, requests and redirects. The maximum redirect count is 0–5. Processing is sequential with a short interval between URLs.

Output

{
  "inputUrl": "https://example.com/",
  "finalUrl": "https://example.com/",
  "statusCode": 200,
  "category": "reachable",
  "method": "HEAD",
  "redirects": [],
  "checkedAt": "2026-09-09T10:54:07.385Z",
  "durationMs": 215,
  "billable": true,
  "message": "HTTP endpoint responds. This does not verify the page content or any claim."
}

This example came from a local, unbilled check. A run summary is stored under SUMMARY in the default key-value store. Output is a snapshot from the time of the run. Re-run when fresh evidence is required.

Category	Interpretation
reachable	Direct HTTP 2xx response
redirected	Redirect chain ends at HTTP 2xx
missing	HTTP 404 or 410
authentication_required	HTTP 401
access_blocked	HTTP 403 or 451; existence unknown
rate_limited	HTTP 429; no retry
server_error	HTTP 5xx; may be temporary
http_error	Other non-success HTTP response
redirect_loop, redirect_limit, redirect_invalid	Redirect chain was not resolved
invalid_input, blocked_target	URL cannot be checked under this tool's limits
timeout, network_error	Result remains uncertain

Current pricing

The live Actor fee is $2 per 1,000 completed URL checks. Platform usage is included for users and is not billed as a separate line item. This is a small usage-priced utility; revenue and payout timing are not guaranteed.

An HTTP response classified as reachable, redirected, missing, authentication required, access blocked, rate limited, server error or HTTP error counts as one completed check. These responses all provide useful evidence and are billable. Input errors, security rejections, unresolved redirects and network failures have no Actor event fee; platform compute costs may still apply. Duplicates do not create additional result rows.

Only the custom url-checked event is charged. Dataset-item and startup charges are disabled. The implementation stops when the platform reports the run's event charge limit.

Automatic resumption is not supported. If Apify restarts or migrates an execution, it stops before replaying checks to avoid duplicate charges within the same run. Results already delivered remain in the dataset; a fresh run performs and charges for its checks again. Disable automatic restart on error for this Actor.

Scope and privacy

Public IPv4 destinations on standard HTTP or HTTPS ports only. Private networks, credentials in URLs, IPv6-only hosts and custom ports are unsupported. Every redirect destination is revalidated. The checked IP is pinned for each request.

No browser, proxies, login cookies, LLM calls or page-body storage. HEAD is tried first; HTTP 405/501 triggers a GET whose body is discarded. Sites can respond differently to a real browser. Bot protection is reported, not bypassed.

Provide only URLs you are authorized to check. Query strings are retained in output, so do not submit private signed URLs, passwords or tokens. The input and results are stored in the requesting account's Apify run storage under its retention settings. Automatic deletion is not implemented.

Local use

Node.js 22 or newer is required. The standalone checker and tests have no external dependencies.

node --test test/checker.test.mjs
node src/cli.mjs examples/INPUT.json examples/LIVE_CHECK.json

For the Apify entry point, install dependencies and run with the Apify SDK's local storage or on the platform. See LAUNCH.md for publication steps.