Redirect Map Builder

https://example.com/2019/06/seo-tips
https://example.com/about-us.html
https://example.com/product.php?id=482
https://example.com/blog/deprecated-tactic

https://example.com/blog/seo-tips
https://example.com/about-us
https://example.com/shop/482-blue-widget

What the results mean

Each row carries a confidence tier and a status. The summary pills tally them:

• Exact (green) — the path is identical, or identical once the file extension is ignored (/about-us.html → /about-us). Trust these.
• Strong (blue) — a matching last slug segment, a shared numeric ID, or a match asserted by one of your pattern rules. Very likely correct; skim them.
• Fuzzy (amber) — the whole path is similar in tokens and scored at or above your accept threshold. Plausible, but review each one.
• Unmatched (red) — no candidate cleared the bar (a below-threshold Weak best is demoted here too). These need a manual pick, content matching, or a 410.
• Already exists (grey) — the old URL is also in your new list, so no redirect is needed; it is dimmed and excluded from the map.
• 410 Gone (purple) — you flagged the page as retired; it exports as a kill-list entry, not a redirect.
• In a loop (purple, left-barred) — chain flattening found the target itself redirects back, which would create a loop. Fix the destination before shipping.

When you run Check destinations are live 200s, each destination also gets a small badge: 200 (clean), N redirects → … (you're mapping to a redirect, not the final URL), or 404/unreachable (a broken destination to fix first).

How it works

The matching engine runs entirely in your browser. Both lists are normalised first — optionally treating /a and /a/ as equal, comparing paths case-insensitively, stripping tracking params (utm, gclid, fbclid), and ignoring file extensions. Then each old URL runs a matcher cascade, strongest first: exact path, exact path minus extension, last-slug match, numeric-ID extraction, and finally whole-path token similarity. The first matcher that fires sets the tier; a below-threshold token best is demoted to unmatched rather than shown as a shaky guess.

The optional trailing-slash normalisation is migration matching only: it lets /a and /a/ represent the same content candidate. The Redirect Chain Mapper keeps them distinct because its job is to expose an actual slash redirect hop.

After matching, a chain-flattening pass detects cases where an old URL points at another old URL and rewrites the map to the final destination (or flags a genuine loop). Correct a row by hand and the tool offers to induce a pattern rule — e.g. /blog/* → /articles/* — and apply it to every similar row at once. The only server calls are opt-in: sitemap import, Tier 2 content matching (fetches title/H1/description only, live for new URLs and from Wayback or Common Crawl for dead old ones), and the destination-liveness check.

Features

• Calibrated confidence tiers (Exact, Strong, Fuzzy, Weak) with a plain reason per row, not a bare percentage.
• A first-class unmatched bucket and a 410 kill-list so retired pages never get silently dropped or dumped on the homepage.
• Per-row override dropdown with alternates, plus Accept all Exact + Strong to clear the easy rows in one click.
• Pattern rules (wildcard or regex) with a live match-count preview, auto-detected directory remaps, and correction-driven rule induction.
• Import from CSV/TXT upload or a sitemap URL; adjustable normalisation and fuzzy-accept threshold.
• Optional Tier 2 content matching and a destination-liveness check that reuses the bulk status checker.
• Exports to Apache .htaccess, nginx, Cloudflare _redirects and Bulk Redirects, and CSV — ordered specific → general, with self-redirect no-ops dropped.
• Autosaves to your browser; download/import the whole project as JSON.

Limitations

It matches on URL structure, not meaning — two unrelated pages that happen to share a slug can match, which is why Fuzzy rows are flagged for review. It does not crawl your site to discover URLs (paste or import them). Sitemap imports follow nested indexes within explicit depth, document, and URL caps. The destination check and content matching are capped per run and only reach public URLs. And it builds the map — it doesn't deploy the redirects; you export the rules and add them to your server or CDN.

Frequently asked questions