Google Photos Is Sending Desktop Backup to the Web—and the PWA Has to Stay Open
A progressive web app has just inherited a job people normally entrust to background desktop software: watching local folders and uploading new files.
As of August 10, Google Drive for desktop no longer supports backing up photos and videos to Google Photos. Google now directs desktop users to the “Back up folders” feature at photos.google.com. They can leave the site open in a supported browser or install the Google Photos web app and keep that running.
This is easy to dismiss as Google rearranging two products. It is more consequential than that. A large consumer service is moving an ongoing, permission-sensitive desktop workflow out of an installed synchronization client and into a browser-delivered application. That makes the change a useful real-world test of how far installable web apps have come—and where their app-like illusion still breaks.
What changed on August 10
Google announced the transition in stages. From June 15, users could no longer configure new Google Photos backup folders through Google Drive for desktop. Existing configurations continued temporarily, but that support ended on August 10.
Photos and videos already uploaded remain in Google Photos. The change affects future uploads from desktop folders, not the cloud library that users have already created.
The replacement workflow lives inside Google Photos. A user opens the service, chooses “Back up folders,” selects local directories, and grants permission for Google Photos to view and upload their contents. Google says files already backed up should not be uploaded again, although an item that was edited or deleted after being backed up may be uploaded.
There is one operational condition that matters enormously: the Google Photos tab must remain open, or the installed Photos web app must be running. Closing the tab stops automatic uploads. Installation therefore improves the windowing and launch experience, but it does not turn the web app into an unrestricted system daemon.
Google also warns that the feature requires a supported, current browser. Its documentation does not promise universal browser availability, and the web-based folder workflow does not support an Apple Photos Library. These are not footnotes for people trusting the product with an important archive; they define the practical boundary of the replacement.
Why this is a meaningful PWA signal
Installable web apps are often marketed through visible features: an icon, a standalone window, offline loading, notifications, or a splash screen. Those details help an app feel at home on a device, but they are not the hardest part of replacing desktop software.
The harder test is whether a web app can handle durable work involving local resources, permissions, long-running activity, recoverability, and clear status. Folder backup touches all five.
By routing this workflow through Google Photos, Google is demonstrating that a browser application can do more than display a cloud library. It can receive access to user-selected folders, inspect their media, identify work to be done, upload files, and report whether backup is preparing, active, paused, or complete. That is a serious application workload, not a dressed-up bookmark.
For founders, this expands the credible territory for web-first products. Media importers, document utilities, creator tools, asset managers, and lightweight synchronization products can increasingly start with a URL and add installation when persistent windowing or easier relaunch matters. A native wrapper is no longer the automatic first answer merely because a product interacts with local files.
But Google’s implementation also shows why product teams must distinguish installable from autonomous. The PWA can continue its work while it is running in its own window, yet users still need to understand that quitting it ends the process. A desktop client sitting in the system tray created a different expectation: invisible, continuous operation after the user walked away.
The real product challenge is trust
Backup software succeeds when users can forget about it without losing confidence in it. That creates a difficult UX target for the web.
A browser permission prompt may correctly protect access to local files, but the product still has to explain what was granted, how long that access lasts, which folders are covered, whether uploads will resume, and what closing the app will do. The platform’s security model and the user’s mental model must line up.
Google’s status indicators—preparing, backing up, complete, and paused—are important because a silent web interface would be unacceptable for this job. Yet status inside the app only helps when people remember to open the app and look. Stronger reassurance may require notifications, launch behavior that is easy to understand, obvious recovery from expired permissions, and warnings when expected backup has not occurred.
A user thread published shortly before the cutoff describes reliability frustrations with the new flow. That report is anecdotal and cannot establish a general defect, but it highlights the standard Google’s replacement will be judged against. For a backup product, occasional uncertainty can be nearly as damaging as a confirmed failure.
Who should care
- Web app founders: This is evidence that mainstream products are assigning consequential local-device workflows to installed web experiences.
- Developers: The implementation is a reminder to test permission persistence, app restarts, duplicate handling, large folders, interrupted uploads, unsupported browsers, and clear recovery paths.
- Product teams: Installation does not erase browser lifecycle constraints. Onboarding must explain when the app needs to remain open and what “background” actually means.
- Investors and partners: The boundary between a website and desktop utility continues to move, potentially reducing the cost of delivering cross-platform tools—but reliability and support costs do not disappear with the native binary.
- Browser makers: If important services adopt this pattern, differences in file access, permission behavior, installation, and long-running execution become commercial compatibility issues rather than developer curiosities.
What changes for installable web apps
The immediate change is not a new standard or browser API. It is market proof: a major provider is willing to make its web application the recommended home for a workflow previously housed in desktop client software.
That shifts the PWA conversation away from whether web apps can copy the surface of native apps. The more useful question is whether they can assume native apps’ responsibilities while remaining understandable, secure, and dependable.
Google Photos offers a mixed but valuable answer. The web app can access approved folders and perform automatic uploads while running. It also remains constrained by browser support, explicit permissions, and an execution model users must not mistake for an always-on background service.
What IndApp watches next
The first thing to watch is whether Google improves resilience and communicates missed or paused backups outside the active Photos window. A background task becomes far more trustworthy when the product can tell users that expected work did not happen.
The second is browser reach. Google’s documentation tells people to switch or update browsers if the option is unavailable, but does not present a simple public compatibility matrix. If the PWA is now a replacement path, supported environments and permission behavior need to be unusually clear.
The third is whether other large services follow. Local folder access is valuable for media libraries, design tools, document processing, publishing, and business workflows. Each successful deployment makes the web a more credible software distribution channel. Each confusing permission failure reinforces the view that serious utilities still need a conventional desktop client.
Google Photos is no longer using its PWA only as a convenient way to open a website. It is asking that web app to carry an ongoing desktop responsibility. The result will be measured not by how native the window looks, but by whether users can trust what happens after they stop watching it.
Further reading
Recent news
