WordPress invalid JSON response: read the save result
You have a page half-written, the block editor throws an invalid JSON response message, and now every instinct says to keep pressing Update until one sticks. Don’t. First, preserve the work already in the editor. Copy the full content, including the title and any reusable blocks, into a local document before another failed save muddies which version WordPress actually retained.

For a WordPress site used to win enquiries in Dubai, the UAE, Liverpool or the wider UK, stalled publishing is more than a minor annoyance. A delayed service page, campaign update or correction to a contact route can leave the business relying on old information for longer than it should.
The message describes a response problem, not necessarily bad content
An invalid JSON response means the editor expected a JSON response from a WordPress REST API save request but received something it could not parse as the required response. The written page may be perfectly ordinary. The failure can instead be an HTML sign-in screen, a firewall challenge, a redirect, a server-generated warning or a valid API error returned in an unexpected form.
WordPress documents the REST API as a foundation of the block editor, with requests and responses represented as JSON objects. That is why a failure at this point can stop an otherwise normal editor save. The WordPress REST API documentation is useful background, but it won’t identify what your particular server returned.
The useful question is not whether the page contains a rogue apostrophe or an unusual character. It is: what did the editor receive when it tried to save?
Protect the draft before testing anything
Save a local copy of the current editor content first. If the page has custom fields, featured media, an excerpt or settings that are awkward to recreate, take a few screenshots as well. This is mundane admin work, but it beats rebuilding a carefully edited page because somebody began clearing caches and changing settings during business hours.
Then make one controlled save attempt. Use the same browser session, note the approximate time, and avoid opening several copies of the page in different tabs. Multiple tabs can create their own confusion around autosaves, revisions and which draft is current.
Capture the one request that failed
Open your browser developer tools and use the Network panel while repeating a single save. Find the request made when Update or Publish was selected. It will commonly relate to a REST API route for the post or page, but capture the actual request rather than guessing its address.
Record the request URL, status code, any redirect chain, response Content-Type, and the response body. Redact cookies, nonces, authorisation headers, personal data and anything that could grant access before sharing it. A screenshot can help, but copied response details are often easier to inspect.
Read the response body before touching permalinks
A status code alone does not settle it. A 200 response can still fail the editor’s JSON check if the body is HTML. The server may have successfully delivered a document, just not the document the editor asked for.
Illustrative example: an editor is working on a new service page. A session expiry or security rule intervenes, and the save request receives an HTML sign-in page with a 200 status. The browser has a successful HTTP response. Gutenberg has received no usable JSON, so it reports an invalid JSON response. Rewriting the page content would achieve precisely nothing.
Sort the result into the right repair lane
Once you have the failed response, the likely cause usually becomes narrower.
- HTML sign-in page or account page: look at session handling, authentication, domain redirects and any security layer applying rules to logged-in requests.
- 403 or challenge page: identify the layer refusing the request, whether that is a host firewall, CDN, security plugin or application rule. A WordPress 403 refusal-layer diagnosis helps separate the blocked route from the person or system responsible for the block.
- HTML error page with a 500 status: investigate the PHP or server error that produced it. A warning emitted before the intended JSON output can be enough to spoil the response.
- JSON error object: this may be a legitimate REST API response describing permissions, validation or another application-level issue. Read the error code and message rather than treating every JSON-related editor error as a routing failure.
- Redirect to another hostname or protocol: inspect the site URL, HTTPS handling, reverse proxy or CDN configuration. The editor needs to reach the expected endpoint while still carrying the right logged-in session.
This is the boundary between a narrowly observed publishing failure and broader REST API diagnosis. If other editor actions, integrations or admin functions are failing too, use the WordPress REST API request triage guide to scope the wider pattern without randomly changing authentication or permalink settings.
Repair the matched cause, not the nearest-looking setting
There is no sensible universal fix for this error. Mass permalink changes, blanket plugin deactivation and disabling authentication controls may alter the symptom while removing useful protections or obscuring the original cause.
If the body is a sign-in screen, fix the login or session path. If it is a security challenge, give the security or hosting owner the exact route, timestamp, response and affected user action. If PHP output is contaminating the response, correct the underlying warning, notice or fatal error in the component producing it. If the response is valid JSON with a permission error, check the affected editor account and the specific REST endpoint capability.
For a business site with several moving parts, this is where a wider WordPress website repair review can be useful. The aim is to establish whether one publishing request is broken or whether a recent hosting, security, PHP or domain change has disturbed several routes.
Verify one ordinary save, then keep the evidence
After the targeted correction, reload the editor and make one small, harmless change. Confirm the update completes, refresh the page, and check that the intended version remains present. If the page matters commercially, check the public URL too, particularly after an update involving redirects, caching or a security service.
Keep the original failed response alongside the successful test details. It gives the next person something better than an editor screenshot and a vague recollection that it stopped working last Tuesday.
If the error persists, or the response points to a server, firewall or session problem you cannot safely alter, send the editor error and redacted save response for a WordPress publishing repair. Include the page type, approximate failure time, response status, Content-Type, redirect details and a redacted body where possible.