I'm working on a macOS app with a Safari web extension. I'm trying to share a SwiftData model between devices using CloudKit synchronization. I am able to get synchronization in the main app on the same device, CloudKit sync works correctly — changes appear in the CloudKit Dashboard under com.apple.coredata.cloudkit.zone. However, in the Safari App Extension, data is saved locally and persists across launches, but never syncs to CloudKit. I have followed the recommended practices for configuring the App Group and entitlements, but the issue persists. Questions: Is there an official limitation preventing Safari App Extensions from connecting to the CloudKit daemon (cloudd)? If not, what entitlements or configuration changes are required for a Safari App Extension to successfully sync with CloudKit? Is the xpc_error=159 from bootstrap_look_up() a known sandbox restriction for this extension type? Any guidance from Apple engineers or others who have successfully used CloudKit from a Safari App Extension would be appreciated. What I’ve confirmed: The extension’s .entitlements includes: com.apple.security.app-sandbox com.apple.developer.icloud-services CloudKit com.apple.developer.icloud-container-identifiers iCloud.dev.example.myapp Same iCloud container ID for both app and extension CloudKit container exists and is initialized in CloudKit Console Running in :Sandbox environment during development Database name in SwiftData matches container identifier (without the iCloud. prefix) The extension’s codesign output shows correct entitlements App Group is configured (although in this case, extension and app use separate stores intentionally) Observed behavior in Console.app logs: CloudKit sync engine initializes in the extension XPC activities are registered for import/export: _xpc_activity_register: com.apple.coredata.cloudkit.activity.export. xpc_activity_set_criteria: ... import. Then a bootstrap lookup fails: failed to do a bootstrap look-up: xpc_error=[159: Unknown error: 159] CloudKit daemon connection error: CKErrorDomain Code=6 "Error connecting to CloudKit daemon" NSCocoaErrorDomain Code=4099 There is no “Will attempt to upload transactions” or “Upload succeeded” logs are ever seen. Symptoms When the extension is run, I see logs like the following in Console.app: [0x13e215820] failed to do a bootstrap look-up: xpc_error=[159: Unknown error: 159] CoreData+CloudKit: -[PFCloudKitSetupAssistant _checkAccountStatus:]_block_invoke(342): Fetched account info for store : (null) Error Domain=CKErrorDomain Code=6 "Error connecting to CloudKit daemon. This could happen for many reasons..."

I'm developing a web extension for Safari on iOS using MV3. The extension is working fine in Chrome, but in Safari I experience some seemingly random issues. I would like to debug it, but here is my problem. I have my iPhone connected via cable to Mac, and it works fine with XCode, so I assume this part is OK. I open Safari or Safari Tech Preview (doesn't matter) on my Mac, developers options are enabled, and in the Develop menu, under my iPhone section, there are things I can debug. There is an entry "[Ext name] - Extension Service Worker" but when I click it, it's empty. Web inspector pops up, but there are no network requests, no logs, nothing. I know the extension is working, because I can stream log to my HTTP server, but I don't see them here at all. I can use console to trigger commands like chrome.storage.local.get(null, console.log) and it shows my local store, so why I don't see any logs? Also, the background script is not visible in the Sources tab, just one weird request: navigator.serviceWorker.register('safari-web-extension://E3449EA7-EC25-4696-8E6C-[ID HERE]/background.js'); </script> Any ideas what went wrong? The entire team of 4 people has the same issue and we can't move forward because of that. Also, the Develop => Service workers or any other menu section doesn't show my service worker. Logs for websites running on my phone are visible and in general web inspector for them works fine.

Summary: Content scripts injected via manifest continue to receive and respond to chrome.tabs.sendMessage() calls even after the user has navigated away from the original page, causing messages intended for the current tab to be handled by zombie contexts from previous pages. Environment: Safari/iOS Version: 18.5 Extension Manifest: Version 3 Expected Behavior: When a user navigates from Page A to Page B: Page A's content script context should be destroyed. chrome.tabs.sendMessage(currentTabId, message) should only reach Page B's content script Only Page B should be able to respond to action button clicks (or other background to content messages). Actual Behavior: When navigating from Page A to Page B: Page A's content script context persists as a "zombie". chrome.tabs.sendMessage(currentTabId, message) reaches zombie context instead of the Page B's one. Hence, it looks like the extension is broken because the content script does not respond to the background messages. Details: Tab ids are properly recognized by both background and content script The problem does not always occur; it occurs on random occasions. It's quite easy to have it reproduced. It can be reproduced easier if user clicks ext icon during site loading (before it fully loaded), triggering ActionClick (ext icon click) event and then sending a msg upon it to the content script Regardless of whether the content script is injected into the tab using manifest.json, registerContentScripts, or executeScript, the problem is still there Once the problem occurs, e.g. user is on macys.com but zombie injected content script believes it's google.com (a previous page), even refreshing the tab doesnt change anything - zombie context is still there (thinking it's still google.com) . Changing a domain to something completely different one could help though. Then going back to macys.com could still lead to the described issue. A zombie content script does not have access to the page's console function and others. Example communication Sending following message from the background to the content script using chrome.tabs.sendMessage() { "tab": { "id": 155, "active": true, "url": "https://www.macys.com/", "title": "Macys.com" } } Results in the content-script zombie context response (the url is taken from the window.location.href) "message": { "type": "ActionClicked", "data": {} }, "response": { "data": { "windowUrl": "https://www.google.com/", "contentReached": true, "timestamp": "1,753,138,945,272", } } }

Hi, I’m working on a Safari extension for macOS, and I’d like the app to use specific system notification settings right after installation. I’m wondering if there’s a way in Swift to programmatically configure the default notification preferences (as seen in System Settings > Notifications > [my app]). Here are the desired settings: Only Desktop – without “Notification Center” or “Lock Screen” Alert Style: Temporary Badge App Icon: Enabled Play Sound for Notifications: Disabled Show Previews: When Unlocked Notification Grouping: Off (I don’t want them to accumulate in Notification Center) Here is the code I’m currently using to display a basic notification: private func handleNotificationRequest(_ message: [String: Any]) { guard let title = message["title"] as? String, let body = message["body"] as? String else { return } UNUserNotificationCenter.current().requestAuthorization(options: [.alert, .badge, .sound]) { granted, error in if granted { self.showNotification(title: title, body: body) } } } private func showNotification(title: String, body: String) { let content = UNMutableNotificationContent() content.title = title content.body = body content.sound = nil // No sound for subtle notification // Create notification that doesn't persist in notification center let trigger = UNTimeIntervalNotificationTrigger(timeInterval: 0.1, repeats: false) let request = UNNotificationRequest(identifier: "fast-url-copy-notification", content: content, trigger: trigger) UNUserNotificationCenter.current().add(request) { error in if let error = error { os_log(.error, "Failed to show notification: %@", error.localizedDescription) } } } OS: macOS 26.0 Thanks in advance, Mateusz

Hi, I’m trying to detect whether my Safari Web Extension is running in Safari or Safari Technology Preview. Is there a reliable way to do that? I can get the executable path of the parent process using proc_pidpath(). However, unlike Chrome or Firefox, Safari extensions run under /sbin/launchd as the parent process, not the responsible process (browser’s binary). In this scenario, I need the executable path of the actual browser process, but I haven’t found a way to get it. Also, Safari doesn’t implement the Web Extension API’s browser.runtime.getBrowserInfo(), unlike Firefox. I haven’t tested it yet, but I’m considering checking the user agent string, though I’m not sure how reliable that would be. Use Case Some users use my Safari extension as a web development tool and want to enable some features exclusively in Safari Technology Preview, while using other features only in standard Safari. If I could detect which browser is in use, I could provide the appropriate functionality for them.

I recently upgraded my device from IOS 18.4 to IOS 26. My web extension has disapeared from safari. I can see it in Settings > Apps > Safari > Extensions and when I turn it on and re-open safari. I just get a mesasge that says "{extension name} is no longer avaiable". I have tried Manifest V2 and Manifest V3 both yield the same results. The current production extension bundled with the IOS app has the same problem. I can no longer use or test my own extension !? Help please !

I'm working on a Safari web extension that uses the nativeMessaging facility to communication with native code. When I want to notify the javascript extension from the embedding application, I use SFSafariApplication::dispatchMessage. As per the documentation, this call ... ensures that Safari is launched and that your extension is running before delivering the message. Everything works fine when the background script is running. However, after the background script gets unloaded at some point in time (non persistent background page, default behavior for a manifest V3 extension), the background script is not reloaded by the message from the native app (background script still appears unloaded in the developer menu of Safari, double-checked using a counter stored in browser.storage.local incremented on message reception). In this case, the completion handler of the application gets no error (error == nil) as if the message was correctly delivered. I was able to reproduce this behavior with the sample app delivered for WWDC20 (after upgrading the manifest from v2 to v3 to make it non-persistent). Is it supposed to work ? What I'm doing wrong ?

I’m a developer working on a Safari Web Extension that’s distributed via the App Store and also tested locally through Xcode. I’m running into an issue that’s affecting my ability to debug errors reported to my Sentry error logging instance from production. The Problem When an error is thrown in one of my extension scripts (e.g., background.js, popup.js, or content.js), the error is sent to Sentry but the captured JavaScript error stack trace replaces the file paths with the webkit-masked-url://hidden placeholder like this: ReferenceError: Cannot access uninitialized variable. at ? (webkit-masked-url://hidden/:14677:28) at ? (webkit-masked-url://hidden/:16307:3) This happens consistently across both App Store builds and local Xcode runs. It prevents me from seeing which script the error came from or resolving the actual source code lines using uploaded source maps in Sentry. My Setup Safari Version: 18.5 (Stable on macOS) Distribution: App Store and local Xcode development Extension Type: Safari Web Extension Error Reporting: Sentry (@sentry/browser SDK) Bundler: Webpack with inline-source-map What I’ve Confirmed I can see the actual source files in Safari’s Web Inspector under the Sources tab when the extension is running. My source maps are uploaded to Sentry correctly and are associated with the matching release. Errors from Safari are being captured by Sentry, but the file URLs are masked, so stack traces cannot be resolved against my original source. My Question Is this behavior (masking file URLs in stack traces with webkit-masked-url://hidden/) intentional for Safari Web Extensions? If so, is there any supported method or workaround to allow exception stack traces to reveal the original script path (e.g., popup.js, background.js) so tools like Sentry or even console logs can point to real locations? I fully understand the privacy/security rationale behind the masking, but as the extension developer, this is making it extremely difficult to debug runtime issues in production. I’d really appreciate any insight into: Whether this masking is expected and permanent behavior If there are any entitlements, debug settings, or Info.plist keys that can alter this behavior for development or for trusted/own extensions If Apple recommends a different way to log extension errors that includes script name or source references Thanks in advance for your help! I’m happy to share more technical details or try out suggestions.

Is there any supported mechanism in Safari Web Extensions (MV3) for capturing or logging network request data (like fetch, XHR, or webRequest) triggered by the web page?

In a Safari Web Extension using Manifest V3, how can a content script access an HTML file that is bundled with the extension (e.g., to inject it as an iframe)? Safari's CSP seem to prevent the use of browser.runtime.getURL() in the MAIN world — is there a recommended way to load such resources securely?

Hi everyone, I’m encountering a critical reliability issue with message passing in my Safari Web Extension on iOS 18.4.1 and iOS 18.5. In my extension, I’m using the standard messaging API. The background script sends a message to the content script using browser.tabs.sendMessage(...), and the content script registers a listener via: browser.runtime.onMessage.addListener(handler); This setup has been working reliably in all prior versions of iOS. However, after updating to iOS 18.4.1 and 18.5, I’ve noticed the following behavior: ✅ The content script is successfully injected, and onMessage.addListener is registered (I see logging confirming this). ✅ The background script sends the message using the correct tabId (also confirmed via logs). ❌ The content script’s onMessage listener is not consistently triggered. ⚠️ This issue is intermittent, sometimes the message is received, sometimes it is silently dropped. ❌ No exceptions or errors are thrown in either script, the message appears to be sent, but not picked up from the content script message listener.

Hi, I’m working with the SFExtensionProfileKey in my Safari Web Extension. As I understand it, this key is to get the UUID of the profile currently in use. However, it seems to be missing (no key in userInfo) when the default profile is active. Also, I haven’t found any API to get a profile’s human-readable name or list all available profiles. Could someone clarify: If the value of SFExtensionProfileKey is absent, can I safely assume the default profile is in use? Is there a supported way to get a profile’s display name? Does Safari expose an API for getting all profiles? Thanks in advance for your insights!

Even default Safari Web Extension project is not displayed on iOS 18.4 simulator ("No extensions installed"), so it's not possible to test extensions in simulator, only on real device.

Hi! I'm working on a web extension for Safari and I need to send messages from the containing application to JavaScript. For this I use the method class func dispatchMessage( withName messageName: String, toExtensionWithIdentifier identifier: String, userInfo: [String : Any]? = nil ) async throws of the SFSafariApplication class. If the site is opened in Safari in normal mode, everything works as expected. However, if the site is "docked", the messages are not transmitted to this "Web App".

I've been unable to successfully get a webpage to send a message to a Safari web extension, no matter what I try doing. I've added the following to my manifest.json file, and it's running manifest v3 { "externally_connectable": { "matches": [ "*://mywebsite.com/*", "*://localhost:3000/*" ] } } My web page executes the following code snippet. I've tried this both while running my site locally (on localhost) and pushed to production. let safariExtensionId = "co.companyname.productname.Extension (ABCD1234)" browser.runtime.sendMessage(safariExtensionId, { greeting: "hello"}, function(response) { console.log("Received response from background page"); console.log(response.farewell); } ); In the Safari web extension's background.js file, I've added the following onMessageExternal listener: browser.runtime.onMessageExternal.addListener((message, sender, sendResponse) => { console.log("Received message from the sender."); console.log(message.greeting); sendResponse({ farewell: "Goodbye!" }); }); This is directly copied from the instructions in this WWDC video: https://developer.apple.com/documentation/safariservices/messaging-between-a-webpage-and-your-safari-web-extension It's also extremely difficult to debug what's happening since the extensions service working frequently does not appear in the Web Extension Background Content menu Is there something I'm doing wrong, or a bug I'm not aware of?

Hi everyone, I’m encountering a serious reliability issue with message passing in my Safari extension on iOS 18.4.1 and iOS 18.5 In my extension, I use the standard messaging API where the background script sends a message to the content scrip. The content script is listening using: browser.runtime.onMessage.addListener(handler); This setup has been working reliably in previous versions of iOS, but since updating to iOS 18.4.1 and iOS 18.5, I’ve noticed that messages sent from the background script are not consistently received by the content script. From my logs, I can confirm that: The background script is sending the message. The content script’s listener is not always triggered. There are no errors or exceptions logged in either script. It seems as if browser.runtime.onMessage.addListener is either not getting registered in time or failing silently in some instances. This issue is intermittent and does not occur all the time. Has anyone else experienced similar issues in iOS 18.4.1 and 18.5? Are there any known changes or workarounds for ensuring reliable communication between background and content scripts in this version? Any help or insights would be greatly appreciated. Thanks!

Hello - we have a Mac application that uses a browser extension and the web extension JS APIs to communicate with Safari. As of macOS 15.4 / Safari 18.4 the tab OnAttached and tab onDetached events are no longer received. After some testing we verified that the events were working properly as of macOS 15.3 / Safari 18.3 but appear to have been broken in macOS 15.4. Note a similar issue was reported previously for Safari 17.6 and was fixed in macOS 15.0 (FB14324177). We have made a TestFlight version of our app (Tabby) available to simplify debugging via https://testflight.apple.com/join/Va8Zdv9d. To reproduce the issue: Install the Tabby TestFlight build on macOS 15.4 or 15.4.1 Open Safari, go to Safari settings and select the Extensions tab Enable the Tabby extension and grant permissions to all windows all the time Open a Safari window with at least 3 tabs Note the open window and tabs displayed in Tabby In Safari, perform a tab detach by dragging a tab out of the window Expected behavior Within Safari the detached tab should now be in it’s own window, and via the onDetached event Tabby should update to show the tab in it’s own window AND removed from the original window. Observed Safari fails to send the onDetached event and Tabby will continue to display the detached tab in its original window in addition to the new window. You can also use the repro steps above to observe the onDetached event being received or not by Tabby in the Safari developer console. The same steps but re-attaching the tab to the original window can be used to observe the onAttached event being received or not. We’ve attached two screen recordings to the Feedback ID below, one showing the events working on macOS 15.3, and one showing the events failing to be received on macOS 15.4.1. Note it also fails on macOS 15.4. FEEDBACK ID: FB17367977

Hi! I'm working on a web extension for Safari and I need to send messages from the containing application to JavaScript. For this I use the method class func dispatchMessage( withName messageName: String, toExtensionWithIdentifier identifier: String, userInfo: [String : Any]? = nil ) async throws of the SFSafariApplication class. If the site is opened in Safari in normal mode, everything works as expected. However, if the site is "docked", the messages are not transmitted to this "Web App". Is it possible to somehow link the container application to the docked website so that messages from the application are received by this "Web App"? That you.

I have an app which has at least two extensions: A Content Blocker extension with a request handler that returns an appropriate NSExtensionItem as part of beginRequest. A different file URL is returned depending upon if the content blocking is on or off by a user setting A Safari Web Extension that includes a toolbar button and popover that enables users to enable or disable the ad blocking of the content blocker extension All three targets (App, Content Blocker appex and Web Extension appex) use an App Group default to read and set the on or off status of the content blocking. When the user changes the content blocking status, the app group default is updated and SFContentBlockerManager.reloadContentBlocker(...) is called. The Content Blocker extension reads the default and then returns the appropriate file URL. The issue is, I have noticed that whenever SFContentBlockerManager.reloadContentBlocker(...) is called from the app, Safari always applies the correct rules from the returned file URL. However sometimes when SFContentBlockerManager.reloadContentBlocker(...) is called from the Safari Web Extension using native messaging, Safari does NOT apply the correct rules from the returned file URL. Using logging I have confirmed that the Content Blocker extension always returns the appropriate file URL irrespective if called as a result of the app or the web extension. Despite this, Safari does not seem to always apply the returned file URL rules when it is called from the Safari Web Extension appex. In these cases, quitting Safari and relaunching it seems to make it apply the rules correctly (obviously this is applying it due to its launch state, not due to the Web extension appex asking it to do so at that point). All targets have access to the App Group location where the active content blocking file URL belongs and the inactive content blocking file URL is within the Safari content blocker target as a resource. I don't think this is a memory status issue as I cannot see the Content Blocker extension being killed when it returns complex rules --- the fact it always works when called via the app also seems to rule this possibility out. This brings up a number of questions: Is calling SFContentBlockerManager.reloadContentBlocker(...) from a different appex, of the same app target and app group supported? (it seems to work sometimes and did work in previous versions of the app). Is there an issue that the Content Blocker extension sometimes returns a file URL that perhaps the calling Web Extension appex may not have access to (even though Safari should via the Content Blocker extension)? Any other ideas of why this may not be working correctly? Has anyone else experienced this? It seems to happen on both iOS and macOS Safari using the same codebase.

We have an existing Safari App Extension distributed outside the App Store (self-distributed). Recently, we converted another browser extension to a Safari Web Extension and used the same bundle ID as the original application to avoid any change on the CX side. After distributing this updated app, we noticed that the Safari extension was disabled on users' machines, and users are now required to manually re-enable it in Safari's preferences. Is this the expected behavior and is there way to avoid this for future updates ?