PWA SEO checklist: get a progressive web app indexed
Short answer Google indexes a progressive web app like any website: it reads each URL fresh, without your service worker's cache or the user's stored data. So the server must send each page's text (server or static rendering), every page needs its own URL with a correct status code, and HTML should come from the network first.
Want help applying this to your business? Talk it through on a free 30-minute call.
Book a callCan Google index a progressive web app?
Yes. A progressive web app (PWA) is a website that can also be installed and work offline. To Google it is just a set of URLs. The app features don't earn ranking, and they don't block it, as long as the content reaches Google.
The catch: Google sees your app as a first-time visitor, every time. Its renderer (WRS, the browser Google uses to run your JavaScript) starts clean on each page:
"WRS does not retain state across page loads: Local Storage and Session Storage data are cleared across page loads." (Google Search Central: Fix search-related JavaScript problems)
Whatever your server sends, plus what your JavaScript builds from it, is what gets indexed. What sits in a returning visitor's cache does not count.
The PWA SEO checklist
| # | Check | Why it matters |
|---|---|---|
| 1 | Each page's main text is in the HTML the server sends (server rendering, static pages or pre-rendering) | Google renders JavaScript later, and many other crawlers never do |
| 2 | Every screen worth finding has its own normal URL, like /services/roof-repair | Google can't use #/ fragment URLs for different views |
| 3 | Links are real HTML links (an a tag with an href address), not click handlers | Google only discovers links built that way |
| 4 | Each URL has its own title, meta description and canonical tag | One shared title makes pages look like duplicates |
| 5 | A missing page returns a 404 status, not 200 | App shells often answer 200 to every path, so error pages get indexed |
| 6 | robots.txt doesn't block your JavaScript, CSS or API files | Google won't render code from blocked files |
| 7 | HTML comes from the network first; only static files come from cache first | Cache-first HTML shows visitors old prices and hours |
| 8 | File names change when their content changes (fingerprinting) | Google caches files hard and may run old code |
| 9 | No content needs a permission prompt, login or stored data | Google declines permission requests and keeps no data between pages |
| 10 | The mobile view has the same content as desktop | Google ranks the mobile version |
| 11 | An XML sitemap lists every public URL, submitted in Search Console | Google says a sitemap helps it crawl "larger or more complex sites" |
Rows 1–6, 8 and 9 come from Google's two JavaScript guides, row 7 from Chrome's service worker documentation, row 10 from Google's mobile-first indexing guide, row 11 from Google's sitemap guide. The sections below explain the ones owners trip on.
What does Google see when it crawls my PWA?
It fetches the URL, reads the HTML, then queues the page to run its JavaScript. Google says the page may wait "a few seconds, but it can take longer than that" (Google Search Central: Understand the JavaScript SEO basics). The same page adds that server-side or pre-rendering "is still a great idea because it makes your website faster for users and crawlers".
Other crawlers are less patient. A December 2024 Vercel study of real crawler traffic found that "none of the major AI crawlers currently render JavaScript", naming OpenAI's, Anthropic's and Perplexity's bots. If your app shell arrives empty, those crawlers see an empty page.
To check one page:
- In Google Search Console, open URL Inspection, enter the URL and run a live test. Look at the rendered HTML and the screenshot.
- In Chrome, right-click and choose View page source. That is roughly what a non-rendering crawler gets. If your service text isn't there, fix row 1.
- Visit a made-up URL, like /xyz123, and check the status code in the browser's developer tools (Network tab). It should say 404.
For the Search Console reports that follow up on this, see Google Search Console for a small business.
Does the service worker help or hurt SEO?
Neither, if it's set up right. A service worker is a script your app installs in the visitor's browser to cache files and work offline. It runs between the page and the network, for people who have visited before.
It doesn't change what Google indexes, for two reasons. Google arrives with no stored state, as quoted above. And even for a person, the first page load isn't handled by the service worker:
"By default, a page's fetches won't go through a service worker unless the page request itself went through a service worker." (web.dev: The service worker lifecycle)
So a service worker can't rescue a page Google can't read. Where it hurts is your customers:
| Pitfall | What visitors see | Fix |
|---|---|---|
| HTML cached first ("cache first") | Last month's prices, hours or staff after you update the site | Network first for HTML; Chrome's caching guide calls network first "great for HTML or API requests" |
| New version stuck "waiting" | Old site until every tab is closed | An "update available, reload" prompt; the new worker waits "until the existing worker is controlling zero clients" (web.dev) |
| Everything cached at install | Slow first visit, wasted data, outdated files | Cache the app shell; add the rest as people browse |
| Server sends the app shell for every path | Google indexes error pages as real pages | Return 404 for unknown paths, or add a noindex tag when the content isn't found |
Speed is where a good cache shows up in search: returning visitors load faster. How much speed counts for ranking is covered in Core Web Vitals and ranking.
Do the manifest, install button or push notifications affect ranking?
No guide from Google Search asks for them. The web app manifest (the file with your app's name, icons and start page) and the install prompt are browser features. Chrome's team says installing needs "a manifest with specific fields such as name and icon"; since Chrome 108 on mobile and 112 on desktop, a service worker is no longer required to install from the browser menu (Chrome for Developers: Revisiting Chrome's installability criteria).
Push notifications need permission, and Google says to "Expect Googlebot to decline user permission requests". Never hide content behind that prompt.
One tool changed: Lighthouse, the audit built into Chrome, no longer has a PWA score. Its April 2024 release notes say Lighthouse "has removed the PWA category". Use its Performance, Accessibility and SEO scores, and URL Inspection for indexing.
Is dynamic rendering still a fix for a PWA?
Only as a stopgap. Dynamic rendering means sending bots a pre-built HTML copy while people get the JavaScript app. Google's page on it now opens:
"Dynamic rendering was a workaround and not a long-term solution for problems with JavaScript-generated content in search engines." (Google Search Central: Dynamic rendering as a workaround)
Google recommends server-side rendering, static rendering or hydration instead.
Example
Illustrative cases with invented numbers.
A cleaning company's booking app (US dollars)
A PWA built with plain React shows 4 of its 30 pages in Google. View page source shows an empty shell. The developer turns on pre-rendering for the 30 public pages, about $2,000 of work, and leaves the booking screens client-side. URL Inspection now shows the service text in the HTML.
A Toronto physio clinic (Canadian dollars)
The clinic raised its first-visit price to C$110, but returning patients still see C$95. The service worker serves HTML cache first. Switching HTML to network first fixes it on each patient's next visit. Google showed C$110 all along, because it never used the cache.
A restaurant with an installable menu app
Every made-up URL returns the menu with a 200 status, and Search Console lists "soft 404" pages. The host is set to return 404 for unknown paths, and the offline page gets a noindex tag.
FAQ
Is a PWA good or bad for SEO?
Neither by itself. It ranks like any site when the server sends readable HTML and each page has its own URL. It fails the same way a JavaScript site fails when it doesn't.
Should my offline page be indexed?
No. It's for visitors without a connection. Add a noindex tag, and keep it out of your sitemap.
Do I need a separate app for Google to find my business?
No. Google indexes a PWA's URLs as web pages, like any other website.
Can I check whether my service worker hides content from Google?
Yes. Run a live test in URL Inspection: Google loads the page without your cache, so what it shows is what gets indexed.
Different for you?
- Your pages aren't in Google at all: website not showing on Google.
- Search Console says "crawled" or "discovered, currently not indexed": pages not indexed.
- You just moved your site to a PWA and traffic fell: traffic dropped after a redesign.
- You'd rather have the build handled: Website Agent.
- You want the pages found and checked every month: SEO & GEO Agent.