WP All in One Migrate is the name that shows up once a WordPress move has already stalled. The file exists, or it doesn’t, and the new server will not take what the old one made. This page sorts the tools by one question: who holds the copy while it is in transit. A later checklist, the one that starts after the plugin says it is done, lives on the migration pitfalls page.
People search a wordpress plugin for migrating site as if that were a product. It isn’t. A wp migrate plugin, wordpress duplicator, and Migrate Guru answer different custody problems. Mixing the names up is how a weekend move turns into a second weekend.
Who is holding the bytes
Every one of these tools has to put the database and the uploads somewhere before the new site can boot. That somewhere is the whole comparison. If the archive sits in your wp-admin, your upload limit and your PHP memory are the gate.
If an installer sits on the new host, the new host’s disk and the file you forgot to delete are the gate. If a vendor’s network holds the copy, their queue and their access to both accounts are the gate.
Price pages and feature grids skip that. They also go stale. What does not go stale is the failure you can see: a timeout, a half-written uploads folder, or a file the public can download.
Tianwen Network, founded in 2017, keeps 30 to 40 export sites running. The plugin arguments that matter are the ones that showed up on those machines, not a vendor’s comparison chart.
The archive that has to pass through wp-admin
All-in-One WP Migration, which is what the query wp all in one migrate is asking for, builds one archive and asks you to import it in the new site’s dashboard. The appeal is obvious. You do not assemble a package by hand, and you do not open a database client.
The cost is the path. The file has to travel through PHP on the way out and on the way in.
That path dies in ordinary ways. The host’s upload cap is smaller than the archive. The request runs longer than the gateway allows. PHP is asked to hold more than the memory limit.
On a 1-core machine with 1 GB of RAM, a memory_limit of 1024M was enough to make three WordPress sites unstable, because the ceiling was larger than the machine. A single PHP process on that class of box measured 80 to 124 MB. Setting the limit to 256M, and cutting the worker count from 6 to 4, is what made the box calm again.
An importer that needs a 1024M ceiling on a 1 GB VPS is not “more compatible”. It is a request the hardware cannot keep.
So the practical test of this plugin is not the logo. Export on the source, note the archive size, and compare it with the destination’s upload cap and its real RAM, not the number in the sales plan. If those three do not fit, a paid extension that lifts the cap only helps when the machine can still finish the request. The extension does not add RAM.
There is a second custody problem. The archive is a full copy of the site, including users, drafts, and whatever a plugin wrote into the database. Importing it “successfully” also imports that.
The pitfalls page walks through the junk people do not mean to bring. Use this plugin when both servers can pass a file through wp-admin and you want one archive, not a remote push.
The installer left on the new host
The wordpress duplicator workflow is a different object. The duplicator plugin builds a package on the source, then you place an archive and an installer on the destination and run that installer in the browser. The new site does not need a working WordPress first. For an empty server, that is the feature.
The matching hazard is the installer and the archive staying on the server after the site boots. A theme directory on one of our own sites once held 36 .backup copies, and every one of them could be downloaded from the public web. They were not a Duplicator package.
They were the same class of mistake: a full copy of something sensitive, left where a URL could reach it. An installer plus a zip in the web root is that mistake with a shorter URL.
Delete those files in the same session as the move, then request the installer URL and expect a 404. If the host’s file manager is the only way you can see the disk, look there too. A plugin that hides its own package from the media library has not hidden it from the web.
Duplicator fits when you can copy two files onto a new account and you will be the person who removes them. It fits badly when the destination already has a live site you must not overwrite, and when nobody owns the ten minutes after the installer says finished.
The push that rewrites the database
WP Migrate, and the paid line people look up as wp migration pro, is built around the database rather than a single downloadable zip. You connect two WordPress installs and push or pull. The part worth paying for, when the URL is changing, is a replace that understands PHP’s serialized strings.
A serialized value stores its own length. A text editor that swaps a domain and leaves the length alone will boot a site that looks half empty. Tables load, options come back blank, and widgets vanish.
The editor did not “corrupt the file” in a dramatic way. It lied about a number PHP trusts.
That is why a find-and-replace in a desktop editor is not a migration plugin. The longer version of that failure is on the pitfalls page. Here the point is narrower.
If the host is staying and the domain or the protocol is what changed, a database-aware push is the tool that matches the job. If you also need every upload, confirm the media option is the one you are running, not the one on the pricing page. A database-only pull is a successful run that leaves every image on the old disk.
If the machine is staying and only the name is changing, the string-level work is moving a WordPress site to a new domain. This page does not repeat that checklist. It only says which tool can rewrite a serialized URL without lying about the length.
Use this family when both sites can be logged into at once, and when the risky part is the URL rewrite. Don’t use it as a way to skip backups. A push can overwrite the destination. Take a dump of the destination first, and store that dump outside the web root.
The copy that never lands on your laptop
Migrate Guru moves the site across a vendor’s network. You install a connector, pick a destination the service knows, and the bytes do not have to pass through your browser. For a large uploads directory, that is the point. Your hotel Wi-Fi is no longer part of the transfer.
The custody shift is the price. A third party holds a complete copy long enough to deliver it. Read what that account can see, and remove the connector when the move is over.
A connector left installed is another admin path. The service’s compatibility list is also the boundary. If the destination is a plain VPS the connector does not speak to, you are back to an archive or an installer, and you have spent the afternoon learning that.
This is the tool for a large file set and a destination the service already supports. It is a poor fit when policy says the database cannot leave your network, even briefly. No plugin setting overrules that policy.
The plugin that only points at one company
A wp engine migration plugin is an on-ramp to WP Engine, not a general mover. It is the right conversation when the destination has already been chosen and the host runs the intake. It answers nothing about a move to a VPS, to another managed brand, or between two servers you already rent. Treating it as a generic wordpress plugin for migrating site is how a team installs three connectors and finishes none of them.
The trade-offs of that host, set beside Kinsta and GoDaddy on the same questions, are on the managed hosting comparison. This page will not repeat them. If WP Engine is not the destination, take their plugin off the shortlist before you compare archive tools.
What a successful plugin run still copies
None of these tools decide whether the site should be copied. They copy the site you have. A product template that shipped a 15 MB image in the first screen made hundreds of product URLs fail Core Web Vitals.
The server was fine. The published targets are an LCP within 2.5 seconds, an INP of 200 milliseconds or less, and a CLS of 0.1 or less, measured at the 75th percentile of real loads. The archive will carry the 15 MB file with perfect fidelity.
Fidelity is not a repair.
A migration plugin also will not fix mail, a cache drop-in aimed at the old host, or a rewrite rule the new server never reads. Those show up after the progress bar hits 100 percent. The page on what breaks after a host move succeeds is the list for that hour.
A platform that will not give you a database at all, such as a hosted builder, is a different job again. The Wix move starts from a URL list, not from a zip.
Copies, doorways, and a ranking promise
A plugin that can clone a site in an afternoon can also clone it onto a row of domains. Google’s spam policies call that out in plain language. Doorway abuse includes “having multiple websites with slight variations to the URL and home page to maximize their reach”, and scaled content abuse is generating many pages to manipulate rankings rather than to help a user.
The policy page is Google’s spam policies. A legitimate move is one site changing address. A fan of near-copies is a different act, and the plugin does not know which one you meant.
The same restraint applies to rankings. Google’s hiring guidance says no one can guarantee a number-one ranking, and it tells you to be wary of anyone who claims a special relationship with Google. That sentence is on Do you need an SEO?
. A cleaner migration can remove self-inflicted errors. It cannot be sold as a position.
When the old dashboard is already gone
If the source no longer logs in, the menu of plugins collapses. A dashboard export needs that dashboard. A remote push needs two live admin accounts.
Migrate Guru needs a connector on a server that still answers. What you have left is the backup you actually stored, or a disk copy the host can hand you.
Restore that copy onto a machine you are allowed to wipe. Then fix URLs with a method that understands serialized lengths. A zip you meant to download, and didn’t, is not a backup.
Check the file is on a disk you control before you cancel the old account. Cancelling first is how a “simple plugin move” becomes a ticket with the old host’s abuse desk.
Write four lines down while the old site still responds. The public URL. The PHP version.
Whether wp-content is in the default place. Where the archive will sit, which should not be a public URL. Those lines take less time than a failed import, and they survive handing the job to another person at 4 p.m.
A twenty-page site and a catalogue are not one job
A short brochure can pass through wp-admin when the upload cap is larger than the archive. A catalogue with years of product images often cannot. The progress bar then stops at the same percentage on every retry, and the plugin gets the blame. The mismatch is the size, not the brand name.
Split the work even when the marketing says one button. Database first, uploads second, URL rewrite third. If this plugin cannot split them, use one that can, or copy the uploads with SFTP outside the plugin and let the plugin handle the database. The plugin is allowed to be the wrong tool for the heavy half.
Can you name the file you will delete, and the URL you will request, before you press the button? If you can’t, you don’t have an operator for this move. You have a licence key. The licence does not stay up to remove the installer.
Pick the tool by custody. Dashboard archive, installer on the new disk, database push, vendor network, or a single host’s intake. Then do the hour the progress bar does not include.
Questions before you pick a migration plugin
Which tool works if the old dashboard is already dead?
None of the dashboard exporters do. You need a backup you already stored, restored onto a server you can wipe, and then a URL replace that understands serialized strings.
Does a finished import keep the old rankings?
No. Google says no one can guarantee a number-one ranking. A clean copy can remove errors you caused. It is not a ranking product.
Why delete the installer the same day?
An installer plus the archive in the web root is a full copy of the site at a public URL. On one of our own sites, 36 backup copies in a theme directory were publicly downloadable. Same class of mistake.
Is a desktop find-and-replace enough when only the domain changes?
No. Serialized values store their own length. A text editor that swaps the name and leaves the length will boot a site with blank options. Use a database-aware replace.