Migrating a WordPress Site from Dev to Production, One Page at a Time
In short: "Dev to production" gets talked about like it's the same problem as a staging sync, but it isn't — a dev site usually isn't a clone of production, it's a separate install with its own domain, its own test content, and often Plain permalinks by default. Promoting it wholesale means dragging all of that along. PostDeploy connects your dev environment and your live site as two independent WordPress installs and pushes one finished post or page at a time, rewriting its media and (on Pro) its Elementor and ACF references to what they should be on the live site — without needing your dev site exposed to the internet at all.
"Dev to production" is a one-way street, not a clone-and-sync
A staging-to-production workflow usually starts with staging being a clone of production — same install, same content, diverging only because of the changes you made on the copy. Dev is different. It's where a page genuinely starts its life: built in Local, DevKinsta, Docker, or whatever runs on your machine, with placeholder images, draft copy, and test data that has no business anywhere near the live site. Treating that relationship like a staging clone — sync the database, reconcile the tables — means production inherits your test content along with the one page you actually meant to ship.
Your dev site doesn't need to be reachable from the internet
This is the part people assume is a blocker and usually isn't. Whichever WordPress install you open the connection from is the one that reaches out — so if you're pushing from your local dev site to a live production site, your dev machine makes an outbound request to production, which is already public and listening. Production never needs to call back into your local environment, so there's nothing to expose, no tunnel to set up, and no port to forward just to send one page live.
The permalink trap most local dev tools hit
A fresh WordPress install — which is exactly what most local dev environments are — defaults to Plain permalinks (?p=123 style URLs), not the "pretty" ones a mature site usually has. A tool that hardcodes /wp-json/your-route/... onto a site's URL simply 404s against that default, which looks like a broken connection when it's really an unannounced assumption about permalink structure. PostDeploy builds every outbound connection URL through WordPress core's own ?rest_route=/... fallback instead, which every WordPress install accepts regardless of its permalink setting — so a brand-new local dev site connects on the first try, not after you remember to switch its permalinks to match production.
What actually happens to the content you push
Push a post or page from dev, and PostDeploy carries its basic post meta, taxonomies, featured image, and the content media it actually references — re-uploaded to production's media library, with the content's image URLs rewritten to point at the new copies, not your dev domain. On Pro, that extends to ACF relationship, gallery, and post-object fields, and Elementor designs — backgrounds, galleries, internal links — resolved to their destination equivalents the same way, so a page built entirely in a local Elementor install doesn't arrive with half its images missing or background colors reverting to a default.
What still needs a manual check
Reference resolution is deliberately scoped to what PostDeploy actually knows it migrated — not a blanket find-and-replace across your dev domain. A plain link typed by hand into a paragraph, pointing at http://mysite.local/some-other-page, isn't part of that pipeline on purpose, since a sitewide string replace risks corrupting serialized data it doesn't understand the shape of. If a page has links like that, the Find & Replace tool cleans them up afterward, scoped only to posts PostDeploy migrated — safer than a raw search-and-replace across the whole database.
A typical connect-once, push-as-you-go workflow
- Add your production site as a connection from your dev install's PostDeploy settings — once, not per page.
- Build and review the page locally, with whatever placeholder data and test plugins your dev setup needs.
- When it's actually ready, push just that post or page to the production connection.
- On Pro, a conflict check warns you if the destination page already changed since your last push, and a snapshot means a bad push has an undo.
wp postdeploy push 214 --to=production
Nothing else on the live site is touched — not its other pages, not content someone else published there since your dev copy was made.
Frequently asked questions
Does my local dev site need to be publicly accessible for this to work?
No. Whichever site you click "Push" or "Connect" from is the one that reaches out — so pushing from your local dev environment to a live production site only requires your dev machine to make an outbound request to production, which is already public. Production never needs to reach back into your local environment.
What if my dev site uses Plain permalinks, like a fresh Local or DevKinsta install?
That's fine. PostDeploy builds its connection URLs through WordPress's own permalink-agnostic ?rest_route= fallback instead of assuming /wp-json/... is reachable, specifically so a freshly scaffolded dev site with Plain permalinks (WordPress's own default) connects correctly from the first push.
Will links pointing at my dev domain get fixed automatically when I push?
Media references get rewritten automatically as part of the push, and Pro also resolves Elementor's internal links and image references to the destination. A plain link hardcoded into ordinary post content, like a URL someone pasted by hand, isn't touched by that pipeline on purpose — use the Find & Replace tool afterward to clean up leftover references to your dev domain, scoped only to posts PostDeploy actually migrated.
Try it
PostDeploy Free connects any two WordPress sites — local dev included — and pushes Posts, Pages, and custom post types between them, Plain permalinks and all. PostDeploy Pro adds ACF, Elementor, and WPML reference resolution, conflict detection, and rollback for when the page is more than plain text and an image. See staging-to-production for the clone-based side of this workflow.