ACCEPTANCE REVIEW / RELEASE TO RELEASE
Review the redirects
that changed.
Compare expected destinations with your before and after URL observations. Find new failures, fixes, and missing evidence before signing off a migration.
Your imported data stays in this browser. This tool compares supplied observations; it does not crawl, fetch, or verify live URLs. No account or payment. Imported content is never sent to analytics. Anonymous page visits go to Vercel; URL query details and fragments are removed.
An early Piko Labs experiment. Free to try. Results depend on your supplied evidence; verify important findings before release.
Tried it? Tell us what worked →1. Bring your evidence
Start with the sample, or import your three CSV files. Nothing is uploaded.
Download sample CSVs: Expected map · Before · After
CSV format and comparison rules
Expected: source_url,expected_url. Each crawl: source_url,final_url,status,redirect_codes. status is the final HTTP response, not the source response. redirect_codes is optional; supply pipe-separated actual redirect response codes, for example 301|302. Use quoted CSV cells for commas or line breaks. Maximum 5 MB and 10,000 rows per file. Duplicate normalized sources are rejected.
Use full http(s) URLs. Browser URL parsing lowercases hosts, removes default ports, resolves dot segments, and serializes URL encoding. Paths remain case-sensitive. Query values, query order, and trailing slashes remain distinct. Fragments and credential-bearing URLs are rejected. A 2xx terminal status and exact expected destination pass. Missing current observations are inconclusive, never resolved. Missing previous observations are newly observed, never regressions. A changed redirect-code sequence is a warning unless you explicitly require permanent redirects. Multiple supplied codes evidence multiple redirects, not a route or loop. Mapping chains/loops are separately labeled mapping-only lint.
Both crawls are evaluated against the same current map. Changed expectations across releases are not compared. Exports include imported evidence; crawl accuracy and freshness remain your responsibility. Remove confidential query parameters before importing.
2. Review the difference
| Source / expected | Result | Previous → current evidence |
|---|
Mapping-only lint
3. Keep the baseline
Save the imported map and both snapshots for your next review. Project JSON is portable and contains URLs. Browser storage is device-specific and can be cleared by the browser.
HELP SHAPE THE NEXT VERSION
Was this useful for a real release?
Tell us what you were checking, whether the report helped, and what you would need to use it again.
Send feedback to Piko ↗Opens your email app with three questions. No imported URLs or CSV data are added. Only send information you choose to share with piko.dream22@gmail.com.
Local event diagnostics · no external transmission
These are prototype events, not unique users, verified customer activity, or revenue. Demo events are labeled and excluded from the imported-data event count. Imported-data events can include tests and repeated use. No user identifiers or imported URLs are recorded.