For web agencies handing over a rebuilt site

Hand over the migration, URL by URL.

Bring two crawl exports and your redirect map. Get back a register that pairs every old URL with its destination, shows what broke, and puts an owner and a next step on each finding.

Free. Your files stay in your browser.

Birch & Beam a fictional studio site: 128 old URLs in file order, each with an example note

Drag or tap the strip, or focus it and press ← →

  • 116 old page over its destination, no exception found
  • 4 blocking: the destination is there but wrong
  • 4 evidence missing or changed
  • 4 blocking: no destination in the files
  • 128 of 128 with a recorded decision

Old URL 1 of 128 · CSV record 2

Blocking

Target is marked non-indexable

owner: Agency developerFollow-up assigned · still open

Current record 2 answers 200 but is marked Non-Indexable. Confirm /services/residential-renovation is meant to be indexed, remove the noindex, then send a corrected crawl for the recheck.

Open this pair in the workspace
FieldOld crawl record 2Current crawl record 2
Address/services/home-renovation/services/residential-renovation
Status code200200
IndexabilityIndexableNon-Indexable
Page titleResidential renovation | Birch & BeamResidential renovation | Birch & Beam

What your client receives

A handover that opens with what is still open.

Not a score, and not 128 green rows. The first page says how many exceptions remain, what happens next and who has it. The follow-up sheet is grouped by owner. Every exception keeps its source rows beside the decision, and all 128 old URLs are listed at the end.

  • handover.html the report, with its fonts, for screen and print
  • decisions.csv one row per old URL, safe to open in a spreadsheet
  • project.json everything, restorable in the workspace
  • originals/ your three files, byte for byte, with SHA-256 identities

Print the report from your browser to save a PDF. The sample is an example of the reviewed handover’s form, written for the fictional project.

Where this handover stands

8 blocking exceptions need follow-up.

128 of 128 original URLs have a recorded decision. Detected exceptions remain open until corrected evidence is supplied.

Next action

Target is marked non-indexable: /services/home-renovation

Current record 2 answers 200 but is marked Non-Indexable. Confirm /services/residential-renovation is meant to be indexed, remove the noindex, then send a corrected crawl for the recheck.

Responsible ownerAgency developer

Uploaded snapshots only. Live redirects, forms, browser rendering, canonical correctness and SEO outcomes are not verified.

10 more open items follow, then all 128 rows.

Two columns of paper cards, lavender on the left and aquamarine on the right, each pair tied by a thread; the last left-hand card is red and its thread hangs loose.

What it flags

Every old URL is paired. The loose ones are flagged.

The comparison reads your files and nothing else. A destination that answers 200 is not cleared by that alone, and a blank cell is reported as blank. In the example, 12 of 128 old URLs raise one of these:

No destination in the map
2An old page with no row in your map has nowhere to go.
Destination missing from the current crawl
2The map points somewhere the new site did not answer for.
Destination crawled in another form
0A trailing slash, letter case or query order apart: named, not called missing.
Destination does not answer 200
2A 404, or a 301 that redirects again.
No usable status code
1A timeout or a blank is not a response.
Answers 200 but is noindex
1The staging setting that follows a site to launch.
Title changed
1Kept on the record even when the change is intended.
Title, H1, canonical or indexability missing
3A blank cell stays a blank, not a pass.
Canonical names another address
1Often the old one, or staging. A relative canonical is flagged too.

Free, in this browser

Try it on your own three CSVs.

Preparing the files or checking the handover? Use the website migration checklist.

Export the old site and the current site from your crawler as HTML-only CSV, and save your redirect map as CSV. Choose the three files and the register is built in this tab. No file is uploaded and no URL in your files is fetched.

The column names are the ones Screaming Frog uses for its internal HTML export. They were checked against a 2021 export; an export from the current version has not been tested yet. If one of your headers differs, rename it to the name in the table before importing.

Example files to copy the format from: old crawl, current crawl, URL map.

The three files
FileRequired columnsRead when present
Old crawl Current crawlAddress, Status CodeIndexability, Indexability Status, Title 1, H1-1, Canonical Link Element 1, Content Type
URL mapOld URL, New URLMerge rationale, required when several old URLs share one destination
Up to 250 rows and 2 MB per file, two origins in total, HTML rows only, UTF-8, comma separated. Every other column is kept and exported unchanged. Sharing for review also has a 16 MiB combined project limit, including saved decisions; it is checked before upload.

The reviewed handover

$249per project · one payment

A reviewed record you can hand over.

Import your three files, then choose to share them for review. Save your private recovery receipt before payment. Turnaround and available capacity are shown before checkout.

No subscription. The free comparison stays available.

One website move, read pair by pair.

What is reviewed

  • Every old URL in your old crawl, against its row in the current crawl and your map.
  • A decision, an owner and a rationale recorded for each pair, written from that pair's rows.
  • A review note, the follow-up sheet and a fixed ZIP you can forward.
  • One recheck. Send a corrected current crawl and map within 14 days of delivery; changed pairs are read again and the first review is kept beside them.

What is not

  • Nobody logs in to your site and nothing is crawled. Only the files you share are read.
  • Live redirects, forms, rendering and rankings are not checked, and no launch is certified.
  • The review is done by an AI agent operator. It is not an independent human expert's audit.
  • Up to 250 HTML URLs per file and two origins per project.

A crawl file is evidence, not a live check.

A 200 in a CSV says the crawler got an answer that day. It does not say the redirect works in a browser, the form submits or the page ranks. Those checks stay with the agency, and the handover lists them as still needed.

A recorded decision never removes a finding. An intended title change stays on the record as a change; a missing canonical stays missing until a corrected crawl shows it.

Your browser keeps the project so you can come back to it. That copy is a convenience, not a backup: download the ZIP or project.json before you clear your browser.