Check whether an old URL reaches the expected new page

Start here — no JSON required

1. Copy two spreadsheet columns (old URL, expected new URL) and paste them into Redirect map. CSV and Excel/Google Sheets tab-separated text both work. A header is optional.
2. Set the run charge limit to cover your small sample, then click Start.
3. Open the results table. Read Result summary and Next step, then export using Download CSV or Download Excel in the output links.

Pasteable example:

old_url,new_url
https://docs.apify.com/api/v2/act-put,https://docs.apify.com/api/v2/actor-put

Recognized headers include old_url/new_url, url/expectedUrl and source_url/target_url. For a single pair, clear Redirect map and fill Old URL and Expected new URL instead. For large migrations, split your map into batches of up to 1,000 pairs. Keep both cells present on every row. Malformed rows stop before requests or result charges rather than being silently dropped.

Sample cost: One pair costs $0.00305 including start; a $0.01 cap is enough to try the sample. 1,000 pairs cost $3.00005; use at least $3.01 for that batch. Unknown/blocked diagnostic results are also charged. Platform usage is included. These examples are independent technical samples, not customers or partners. To use Advanced JSON, clear the beginner text fields; do not fill two input methods at once. Existing API inputs remain supported.

Only treat results as complete when the run succeeded and OUTPUT.status is complete. An unknown result needs review; it is not a verified negative. Start small: network timeouts and access blocks can still occur.

After a website migration, an old URL returning 200 is not enough: it might land on the homepage or the wrong article. Supply old/new URL pairs and compare the actual redirect destination with your intended target.

Input

{
  "checks": [
    {
      "url": "https://docs.apify.com/api/v2/act-put",
      "expectedUrl": "https://docs.apify.com/api/v2/actor-put"
    }
  ]
}

The example is an independent check of a public documentation redirect. Replace both URLs with your own migration map. Batches support up to 1,000 pairs; split larger maps across runs.

What a result means

Each row contains the requested URL, expectedUrl, finalUrl, HTTP statuses, full observed chain, redirectCount, matchesExpected, canonicalUrls, robots directives, issues and a verdict.

• pass: one permanent server redirect (301 or 308), the expected final URL, a 2xx response, one matching canonical, no observed noindex and no detected checks requiring review.
• review: destination checks passed but there are cautions such as extra hops, a temporary redirect, no observed canonical, multiple canonicals or truncated HTML.
• fail: a verified wrong destination, missing source redirect, redirect loop, unhealthy destination, HTTPS downgrade, observed noindex or mismatched canonical.
• unknown: a request could not be verified, including bot blocks, authentication, rate limits, server errors, networking errors or the redirect-hop limit. Do not count these as confirmed broken links.

URL comparison retains query parameters and trailing slashes, removes fragments, normalises hostname case and omits default ports. It does not equate /page and /page/ or reorder query parameters. Supply the intended final URL exactly.

Static HTML canonical links and robots/googlebot meta directives, plus X-Robots-Tag, are diagnostic hints. This is not a Google indexability certification. The tool does not render JavaScript, follow meta refresh, inspect robots.txt, find soft 404s, assess content equivalence, generate redirect rules or modify your website. A 200 challenge page without a detectable status block may need manual review. HTML beyond 2 MB is truncated and cannot receive a clean pass.

Pricing and completion

$3 per 1,000 URL check results ($0.003 per dataset row), plus $0.00005 per run start at the supported memory settings. Platform usage is included. Each supplied check produces one billable diagnostic row, including blocked, unavailable, invalid and uncertain results. This price is per check, not per detected change. A 10-check completed run costs $0.03005 at these prices.

Use a small sample first. Set the maximum run charge to cover all checks plus the start fee. Before fetching URLs, the Actor checks the budget against the run's active price snapshot. If insufficient, it fails with no URL requests and no result records; the start fee can still apply. A storage failure after a dataset write may leave charged partial results. Writes are not retried. Start a fresh run rather than resurrecting a partially delivered one.

Only treat the export as complete when the run succeeded and the OUTPUT key-value record has status complete. SUMMARY contains counts; dataset rows can be exported through Apify. Completion means every requested diagnostic row was delivered, not that every remote page was reachable or passed its checks.

Limits and handling

1–1,000 checks per run; 1 MB input; 256 MB default memory, 512 MB maximum; 3,600-second default run timeout. Public HTTP/HTTPS on standard ports only; up to five redirects and 2 MB of response content. Static HTTP requests only: no browser rendering, login, proxy, CAPTCHA handling, or retries. Private/network-local destinations are rejected on every hop, and DNS answers are pinned to the connection. URLs and results remain in the initiating account's Apify stores under its storage/access policies. The Actor contacts supplied public websites and its own Apify stores; it sends no user credentials to target sites. Use non-confidential URLs.

An early product, tested with synthetic edge cases and a small public sample. No customer time savings, revenue, or complete website coverage are claimed. For support, use the Actor's Issues tab with a non-confidential input example, expected result and actual result.