On this page
- Headless CMS SEO: what changes when the CMS has no front end
- Our headless SEO services
- Rendering for headless SEO: SSR, SSG, ISR or client-side
- Content modeling for headless SEO
- Metadata, canonicals, hreflang and sitemaps your code generates
- Speed, images and Core Web Vitals on a headless front end
- Preview and staging leaks
- Structured data from the CMS model
- AEO and GEO: AI crawlers that don't run JavaScript
- Migrating to a headless CMS without losing traffic
- Choosing a headless CMS, and an agency for it
- Headless CMS SEO: FAQ
Rankstar is a headless CMS SEO agency for sites whose content lives in Contentful, Sanity, Strapi, Storyblok or an in-house CMS and whose pages are built by Next.js, Nuxt or another JavaScript front end. A dedicated consultant checks what your code actually serves to Google and AI crawlers, specifies the SEO fields your content model needs, writes content, builds links and tracks your visibility in Google AI Overviews, ChatGPT, Gemini and Perplexity. See the plans or book a call.
Headless CMS SEO: what changes when the CMS has no front end
In WordPress or Shopify, the platform prints the title tag, canonical, sitemap and redirects for you. In a headless setup the CMS only stores content and serves it through an API. Your front-end code produces every SEO signal, and anything a developer didn't build simply isn't there.
Your code owns:
- Rendering: whether the HTML a crawler receives contains the content.
- Title, meta description, Open Graph and robots tags.
- Canonical URLs, trailing-slash rules and HTTP status codes (200, 301, 404, 410).
- Hreflang for each locale.
- XML sitemaps and robots.txt.
- Redirects after slugs change.
- Structured data, image markup and internal links.
So is headless better for SEO? It can be faster and gives you full control of every tag, but the same freedom means nothing comes by default. Headless SEO is mostly about making sure that control is used, template by template.
Our headless SEO services
Each month of our SEO, AEO and GEO packages covers:
- Technical audit (month 1). We compare the raw server HTML with the rendered page for each template: titles, canonicals, hreflang, status codes, sitemap coverage, structured data, internal links and Core Web Vitals. Each finding comes as a ticket your developers can pick up.
- An SEO field spec for your content model. The fields, validations and fallbacks editors need, written for Contentful, Sanity, Strapi, Storyblok or your own schema. On an in-house CMS we first document which SEO signals it outputs today, because that list rarely exists.
- Keyword research and a topical map. Which content types and pages to build, mapped to search intent.
- Content. 1,000+ word articles and page copy each month by native US, UK or French writers, split into your fields. We write, you publish.
- Off-site placements. Hand-picked links through PR, guest posts and niche edits, brand mentions in "best of" listicles and reviews, and Quora answers.
- Google Autosuggest from the Essential plan up, and a dashboard updated monthly from your Google Analytics and Search Console.
Deliverables arrive within 28 days, or 14 with express delivery. Compare the plans or book a call.
Rendering for headless SEO: SSR, SSG, ISR or client-side
Google's JavaScript SEO basics describe three phases: crawling, rendering and indexing. Rendering can happen later than crawling, so content that only appears after JavaScript runs reaches the index later, or not at all if a script fails. Google calls server-side rendering or pre-rendering "still a great idea" and notes that not all bots can run JavaScript. It also describes dynamic rendering, serving bots a separate pre-rendered version, as a workaround rather than a long-term solution.
| Mode | What a crawler gets on the first request | SEO notes |
|---|---|---|
| SSR (server-side rendering) | Full HTML built per request | Safest for content that changes often; watch server response time. |
| SSG (static generation) | Full HTML built at deploy | Fast and cacheable; new or edited entries wait for the next build. |
| ISR (incremental static regeneration) | Full static HTML, refreshed on a timer or webhook | Good middle ground; connect CMS publish webhooks so edits go live quickly. |
| CSR (client-side rendering) | A near-empty shell plus JavaScript | Google may render it later; most AI crawlers won't. Avoid for anything you want found. |
Next.js App Router pages render on the server by default, and Nuxt uses universal rendering by default, so the usual problem isn't the framework. It's main content fetched in the browser after load, product data behind a client-only component, or links built as click handlers instead of real a href elements. Google also asks for the History API for routing, not URL fragments, so every page has a crawlable URL.
Content modeling for headless SEO
The content model is where headless SEO is won or lost before a page exists. We look at three things:
- URL structure from the model. Slugs, parent pages and categories should produce short, stable, readable URLs, with the path rules written once in code.
- Topics as content types. When categories, topics, authors and products are entries rather than text fields, the front end can build hub pages, breadcrumbs and related links automatically. That is how internal linking scales.
- One canonical home for reused content. Omnichannel delivery means the same entry can appear on the website, in an app and in partner feeds. Pick the web URL that should rank and point any duplicate web copies to it with a canonical.
Metadata, canonicals, hreflang and sitemaps your code generates
SEO fields in the content model
Every page-level content type needs an SEO title, meta description, slug, canonical override, noindex switch and social image. We add length validation in the CMS and fallbacks in code (for example, the entry title when the SEO title is empty), so a missing field never ships a blank tag.
Canonicals and status codes
Put an absolute canonical URL in the server HTML. Google says it's possible to inject one with JavaScript but the best way is HTML. Pick one trailing-slash rule and redirect the other. Unknown slugs must return a real 404, not a 200 page that says "not found": headless front ends produce these soft 404s more than any other setup we audit.
Hreflang
Build hreflang from the CMS locale data, with each version listing every other version and itself. Google accepts it in HTML link tags, HTTP headers or the sitemap; one method done consistently beats three done halfway. Translated slugs need their own redirect rules.
Sitemaps and redirects
Generate sitemaps from published entries only, with lastmod taken from the entry's real update date. Google caps a sitemap file at 50,000 URLs or 50 MB uncompressed, so large sites need a sitemap index. Store redirects as a CMS content type or in your edge configuration, and serve them as server-side 301s so a changed slug never breaks old links.
Speed, images and Core Web Vitals on a headless front end
Headless sites are often sold as fast, and they can be, but the front end decides. Google's Core Web Vitals measure loading (LCP), responsiveness (INP) and visual stability (CLS), and the causes we find on headless builds are specific:
- Images. Request the right width and a modern format from the CMS image API, serve responsive sizes, load the main (LCP) image eagerly and lazy-load the rest.
- JavaScript weight. Large client bundles and heavy hydration slow responsiveness; keep interactive components small and render everything else on the server.
- Server response. SSR pages that wait on several API calls start slowly; cache CMS responses and serve static or ISR pages from the edge where the content allows.
- Fonts and third-party scripts. Self-host fonts, reserve space for embeds so the layout doesn't shift, and load tag managers after the main content.
Preview and staging leaks
Headless stacks multiply URLs: preview deployments for every branch, a staging domain, CMS preview links that show drafts, and sometimes the content API itself. When those get indexed, they compete with production pages and can expose unpublished content.
- Password-protect non-production environments. Google's page on controlling what you share says private content needs password protection; noindex only keeps a page out of results while it stays reachable.
- Make robots rules depend on the environment. Production allows crawling; every other environment sends noindex or blocks access.
- Never ship noindex in the server HTML and remove it with JavaScript. Google may skip rendering a page once it sees noindex, so the page stays out.
- Keep drafts out of sitemaps and feeds. Filter on published status, not on whether the API returned the entry.
Structured data from the CMS model
A headless CMS already holds structured content, so JSON-LD can come straight from the fields: Article with author, datePublished and dateModified from the entry timestamps; Product and Offer from catalog data; BreadcrumbList from the route tree; Organization once, sitewide. Render it on the server with the page, and keep it in step with what the page shows: Google lists structured data that matches the visible text as a best practice for its AI features. Since August 2023 Google shows FAQ rich results only for well-known government and health sites, so FAQPage markup is there to describe the page, not to win a snippet.
AEO and GEO: AI crawlers that don't run JavaScript
This is where headless SEO most often falls short. A December 2024 study by Vercel and MERJ found that none of the major AI crawlers rendered JavaScript, including OpenAI's crawlers, ClaudeBot and PerplexityBot. Gemini uses Googlebot's infrastructure and AppleBot renders pages, but a client-rendered page that Google indexes can still look empty to the crawlers behind ChatGPT, Claude and Perplexity.
- Server-render the main content, metadata and navigation of every page you want cited.
- Decide in robots.txt which AI crawlers you allow: OpenAI separates GPTBot (model training) from OAI-SearchBot (ChatGPT search), and Google-Extended controls whether Google uses your content for Gemini training and grounding, without affecting Google Search.
- Write answer-first sections, comparison tables and clear definitions that an assistant can quote.
Google says AI Overviews and AI Mode need no special files or markup beyond indexable pages eligible for a snippet. The rest of GEO happens off your site: for prompts like "best headless CMS for ecommerce", assistants cite listicles, reviews and forum threads, so we work on placing your brand there through AI brand mentions and Quora answers. Google Autosuggest work gets your brand suggested as people type.
Migrating to a headless CMS without losing traffic
The SEO considerations when migrating to a headless CMS come down to parity: the new site must serve at least what the old one did. Moves from WordPress, Drupal (see our Drupal SEO guide) or a legacy custom CMS fail when the old platform was doing SEO work nobody listed.
- Crawl the old site and record every indexable URL with its title, canonical, hreflang, structured data and internal links.
- Export the URLs with clicks or backlinks from Search Console and map each to its new URL.
- Build the new templates against that inventory, then compare rendered HTML on a password-protected staging site.
- At launch, check status codes, robots.txt, canonicals and the new sitemap, and watch Search Console coverage daily.
Google's site move guidance asks for permanent server-side redirects (301 or 308) kept for at least a year. Our SEO audit can run before launch as a parity check.
Plans for headless and custom CMS sites
Every plan includes a dedicated consultant, a technical audit of your rendered pages, monthly content and off-site placements.
-
Launch
- 1 × Technical audit
- 1 × Keyword research
- 1 × Topical map
- On-site SEO / AEO / GEO
- 10 focus SEO / AEO / GEO keywords
- 2 × HQ content (1,000+ words)
1 off-site placement
- 1 × HQ link
- White-label report
- Tracking dashboard, updated monthly
- Dedicated consultant
-
🔥 Popular
Essential
- 1 × Technical audit
- 1 × Keyword research
- 1 × Topical map
- On-site SEO / AEO / GEO
- 20 focus SEO / AEO / GEO keywords
- 5 × HQ content (1,000+ words)
2 off-site placements
- 1 × HQ link
- 1 × AI citation (listicle)
- 1 × Autosuggest keyword
- White-label report
- Tracking dashboard, updated monthly
- Dedicated consultant
-
Growth
- 1 × Technical audit
- 1 × Keyword research
- 1 × Topical map
- On-site SEO / AEO / GEO
- 50 focus SEO / AEO / GEO keywords
- 10 × HQ content (1,000+ words)
4 off-site placements
- 2 × HQ links
- 1 × AI citation (listicle)
- 1 × AI citation (review)
- 2 × Autosuggest keywords
- White-label report
- Tracking dashboard, updated monthly
- Dedicated consultant
-
Pro
- 1 × Technical audit
- 1 × Keyword research
- 1 × Topical map
- On-site SEO / AEO / GEO
- 150 focus SEO / AEO / GEO keywords
- 20 × HQ content (1,000+ words)
7 off-site placements
- 3 × HQ links
- 2 × AI citations (listicle)
- 2 × AI citations (review)
- 8 × Autosuggest keywords
- Google AI Overviews secret sauce
- White-label report
- Tracking dashboard, updated monthly
- Dedicated consultant
-
Authority
- 1 × Technical audit
- 1 × Keyword research
- 1 × Topical map
- On-site SEO / AEO / GEO
- 250 focus SEO / AEO / GEO keywords
- 30 × HQ content (1,000+ words)
20 off-site placements
- 4 × HQ links
- 3 × AI citations (listicle)
- 3 × AI citations (review)
- 10 × Quora answers
- 10 × Autosuggest keywords
- Google AI Overviews secret sauce
- White-label report
- Tracking dashboard, updated monthly
- Dedicated consultant
Choosing a headless CMS, and an agency for it
People often ask for the best headless CMS for SEO. Most headless CMS platforms for JavaScript SEO can do the job, because the front end decides most of the outcome, so judge the CMS on what it lets your developers and editors do:
- Field validation and required SEO fields per content type.
- Preview through a draft mode that stays out of the index.
- A localization model that maps cleanly to hreflang.
- Publish webhooks to trigger rebuilds or revalidation.
- An image API that serves modern formats and sizes.
- A way to manage redirects as content.
For the agency, look for one that reads rendered HTML rather than only running a tool, writes specs your developers can ship, covers AI crawlers and off-site mentions, and shows real reporting (see our case studies). Most headless builds we see are software and commerce sites; our SaaS SEO agency and ecommerce SEO agency pages cover those markets, and our guide to SEO for SaaS companies goes further.
Running part of your stack on a hosted platform? We also work on Webflow, WordPress, Shopify, Magento, BigCommerce, WooCommerce, PrestaShop, Wix and Squarespace.
Headless CMS SEO: FAQ
Is a headless CMS good for SEO?
It can be. A headless CMS gives you full control of rendering, metadata and speed, but nothing comes by default: titles, canonicals, sitemaps, redirects and structured data all have to be built in the front end. Sites that build them well often outperform template platforms.
What is the best headless CMS for SEO?
There isn't one winner. Contentful, Sanity, Strapi and Storyblok can all support strong SEO. Choose on field validation, draft preview, localization, publish webhooks, image handling and redirect management, then make sure the front end renders content on the server.
Does Google index JavaScript sites?
Yes. Google crawls, renders and indexes JavaScript pages, but rendering can happen later than crawling and some scripts fail. Google still recommends server-side rendering or pre-rendering, and many other crawlers, including most AI crawlers, don't run JavaScript at all.
Which headless CMS SEO tools do we need?
Start with Search Console's URL Inspection, which shows the HTML Google rendered. Add a crawler that can render JavaScript to compare raw and rendered pages, PageSpeed Insights for Core Web Vitals, and Google's Rich Results Test for structured data.
Is a headless CMS better for SEO or AI search?
Neither by default. A server-rendered headless site serves clean HTML that both Google and AI crawlers can read. A client-rendered one may rank in Google yet stay invisible to ChatGPT and Perplexity crawlers. Rendering, not the CMS, makes the difference.
Do you work directly with our developers?
Yes. Technical findings arrive as tickets with the template, the problem, the fix and how to test it, and content arrives split into your CMS fields. We write and specify; your team ships and publishes. The SEO audit is a good starting point.
What should we check before migrating to a headless CMS?
Record every indexable URL with its title, canonical, hreflang, schema and links, then map each old URL to a new one with a 301 redirect. Compare rendered HTML on a protected staging site before launch and watch Search Console daily afterwards.
How do we keep preview and staging sites out of Google?
Password-protect them. Google says password protection is how to keep private content out of search; noindex alone leaves pages reachable. Make robots rules depend on the environment, and keep drafts out of sitemaps by filtering on published status.
Make your headless site readable to Google and AI
Send a 5-minute brief after checkout. Your consultant starts with a technical audit of your rendered pages, keyword research and a topical map, then delivers content and placements every month.











