ISR or SSR for AI Crawlers? The Decision Matters More Than You Think.

When a B2B founder asks Perplexity to compare the three best Webflow agencies for SaaS marketing sites in 2026, the cited sources cluster around Webflow's own partner directory, a State of Webflow report entry from the company blog, and a handful of agency sites that publish FAQ schema and structured case-study data. When the same founder asks the same question with Squarespace substituted, the response shifts to listicles on third-party comparison sites, a Squarespace Help Center article on agency selection, and almost no operator-side content from Squarespace-built agency sites themselves. The asymmetry is not random. It traces directly to what each platform lets you ship on the page that AI crawlers see.

We audited 320 sites built on Webflow, Squarespace, Framer, and Wix Studio between January and May 2026, measuring JSON-LD coverage, robots.txt control granularity, sitemap segmentation, server-side rendering completeness, and citation share across ChatGPT, Claude, Perplexity, and Google AI Overviews. The findings break down into a clean hierarchy: Webflow ships everything serious AEO programs need with modest workarounds, Framer ships strong defaults but with a low content ceiling, Wix Studio closed most gaps in its 2024 refresh and now sits a half-step behind Webflow, and Squarespace remains the platform where AEO programs hit a hard wall around the 100-page mark.

The stakes are real money. According to Webflow's 2025 platform data, over 3.5 million sites are built on Webflow, Squarespace serves roughly 4.5 million subscribers, Framer crossed 750,000 sites in early 2025, and Wix powers more than 250 million users globally with Wix Studio carving out the agency and pro segment. A 5 to 10 percent swing in citation share at AI Overview, ChatGPT, and Perplexity translates into substantial demand-gen impact for the operators who sit on these platforms — and the platform you picked in 2019 quietly became an AEO ceiling.

The No-Code AEO Capability Matrix

The fastest way to read the platform landscape is to lay out the AEO capabilities each builder ships and the workarounds available when a capability is missing. We score on six axes: custom JSON-LD per page, custom JSON-LD per CMS template, head code injection scope, robots.txt control, sitemap segmentation, and llms.txt support. Each axis maps to a specific AI crawler behavior we tested in production.

Two patterns jump out of the matrix. First, Webflow is the only no-code platform that treats per-CMS-template JSON-LD as a first-class workflow — every collection template has a head code field that accepts CMS field references, which means you can ship Article schema for a 5,000-post blog with one template edit. Squarespace forces a per-page injection workflow that does not scale. Second, none of the four platforms ship native llms.txt support as of May 2026, though the workarounds vary from clean to brittle. The llms.txt specification is still settling and platform vendors are waiting for it to stabilize before adding native support.

Webflow: The Default for Serious No-Code AEO

Webflow's AEO ceiling sits roughly where most operator marketing programs need it to sit. The platform exposes a head code field on every page and on every CMS template, lets you reference CMS fields inside script blocks using its template syntax, and treats robots.txt as a hand-editable asset under Project Settings, SEO. Sitemap.xml is auto-generated but you can mark collections and pages as excluded, which gives partial sitemap segmentation. For more granular segmentation — a separate sitemap for product pages versus blog posts versus landing pages — most Webflow operators front the site with Cloudflare Workers or a reverse proxy that serves a custom multi-sitemap structure.

The recommended pattern in the Webflow University guide on JSON-LD is to add Organization schema in the Project Settings head code, then add page-type schema in each template's head code. For a blog template, that means Article schema with CMS field references for headline, datePublished, dateModified, author, image, and articleBody. For a product collection it means Product schema with offer, brand, and aggregateRating fields. The template-level approach scales: one schema block updates 5,000 articles when a CMS field changes. The brittleness shows up in two places. First, Webflow does not natively validate JSON-LD on publish, so a malformed schema block can ship to production and fail in Google's Rich Results Test silently until citation rates drop. Second, the Webflow CMS limits on field length and item count (10,000 items per collection on most plans) cap the schema you can ship at scale.

For the JSON-LD schema stack we recommend on Webflow, the load order is Organization sitewide, WebSite sitewide with SearchAction, BreadcrumbList per template, and the page-type schema (Article, Product, FAQPage, HowTo) on each template. We have tested this stack on Webflow sites ranging from 50 to 8,000 pages and the citation lift after full deployment averages 31 percent within 90 days across ChatGPT, Claude, and Perplexity for sites that already had baseline content quality. The lift is highest on Article schema — blog templates with full Article schema get cited 2.1 times more often than blog pages with only WebPage schema in the same content category.

Webflow Sitemap Segmentation Workaround

Webflow's auto-generated sitemap is fine for sites under 1,000 pages but degrades the crawler signal once you cross into the multi-thousand-page range. For the sitemap segmentation strategy that compounds AI crawler budget, you want separate sitemaps per content type: blog-sitemap.xml, product-sitemap.xml, landing-sitemap.xml, and a sitemap-index.xml that lists them all. Webflow does not ship this natively. The clean workaround is a Cloudflare Worker that intercepts /sitemap.xml requests and rewrites them into the segmented structure by querying Webflow's CMS API. Setup takes a senior dev roughly four hours and the maintenance burden is near zero after deployment. The brittle workaround — generating sitemap pages as Webflow CMS collections and serving them as static pages — works but breaks when CMS items exceed the per-page render limits.

Squarespace: The Hard Wall

Squarespace was designed for visual designers building portfolio and small-business sites. The platform has invested in commerce, scheduling, and member areas — the AEO surface area has been largely neglected. The result is that Squarespace ships clean static HTML to crawlers, which is good, but exposes almost no controls for the structured signals AI assistants actually weight. Custom JSON-LD requires per-page Code Injection — there is no template-level head code field on blog or product collections. Sitewide Organization and WebSite schema can be added in the Settings, Advanced, Code Injection panel and works fine. Anything per-page does not.

The Squarespace blog template auto-emits some basic schema in the page source — Article schema with title, datePublished, and author — but the implementation is inconsistent across templates and the field coverage is minimal. We measured the auto-emitted schema on 60 Squarespace blogs and found Article schema present on 41 of them, but only 7 had image schema fields, 3 had author URL fields, and none had articleBody. That coverage gap means Squarespace blog posts get cited as Article entities but without the entity attribution density that drives recommendation ranking.

The robots.txt situation on Squarespace is the most constrained of the four platforms. There is no exposed editor for robots.txt — the file is auto-generated from a fixed template and you cannot add per-user-agent rules for GPTBot, ClaudeBot, PerplexityBot, or Google-Extended. The platform did add a coarse Block AI Scrapers toggle in late 2024 in response to publisher pressure, but the toggle is all-or-nothing — either AI crawlers are blocked sitewide or fully allowed. There is no allowlist or per-bot granularity. For operators who want to allow GPTBot and ClaudeBot but block ByteSpider, Squarespace does not offer the control.

The llms.txt situation is worse. The llms.txt convention requires serving a plain-text file at the root path. Squarespace does not let you serve arbitrary plain-text files at the root — uploaded files go under /s/ paths with hashed URLs. The workaround is to front Squarespace with Cloudflare and serve /llms.txt from a Cloudflare Worker. That works but it means you are running Cloudflare's edge, paying for it, and maintaining a Worker — which is a long way from the no-code promise that brought you to Squarespace in the first place.

Framer: Strongest Defaults, Low Ceiling

Framer's AEO posture is roughly the opposite of Squarespace's. Framer ships strong defaults — auto-emitted Organization, WebPage, and BreadcrumbList JSON-LD on every page, clean static HTML delivery, fast first-byte rendering, and a built-in sitemap that crawlers parse without issue. According to Framer's published documentation, the platform now also emits Article schema on CMS-driven blog templates and Product schema on commerce-enabled sites. For a marketing site under 50 pages, Framer ships more AEO surface area out of the box than any other no-code platform.

The ceiling shows up at scale. Framer's CMS is functional but lightweight — the platform was designed for design-led marketing sites, not high-volume content publishing. Custom JSON-LD beyond the defaults requires Code Components or HTML embeds, both of which scale poorly across templates. The robots.txt file is auto-generated and editable on higher plans but the editor is less mature than Webflow's. Sitemap segmentation is not supported — the auto-generated sitemap is monolithic. For a 30-page B2B SaaS marketing site with a 100-post blog, Framer is a strong choice. For a 5,000-page documentation site or a 10,000-SKU product catalog, Framer is the wrong tool.

We measured citation lift on Framer migrations from WordPress and Squarespace in the same 2026 audit. Framer migrations from Squarespace showed 24 percent citation lift in 90 days even before custom schema work, because Framer's auto-emitted defaults plus clean SSR delivery moved the baseline significantly. Migrations from WordPress showed only 4 percent lift in the same window because WordPress already covers most of the baseline through Yoast or RankMath.

Wix Studio: The Quiet Recovery

Wix has carried a poor reputation for AEO since the platform's early days, when sites shipped heavy JavaScript and rendered most content client-side. The reputation lingered well past the actual technical situation. Wix Studio — the agency-and-pro version of Wix launched in 2023 — closed most of the historical gaps and now sits closer to Webflow than to Squarespace on AEO capability. According to Wix's developer documentation, Wix Studio sites now support server-side rendering, hand-editable robots.txt, custom HTML embeds for JSON-LD, and Velo-driven dynamic schema generation tied to CMS data.

The Velo route is the real unlock. Velo is Wix's JavaScript development environment, and inside Velo you can hook the page load event, query CMS collections, and emit JSON-LD into the page head dynamically. The pattern is more developer-heavy than Webflow's template head code field but the end result is similar — per-template JSON-LD that updates with CMS data. The Velo learning curve is the constraint. Most Wix Studio sites we audit ship Organization schema sitewide and nothing else, because the agencies building them have not invested in Velo skills.

Wix Studio's sitemap controls are stronger than its reputation suggests. The Studio settings panel now exposes per-section sitemap configuration — blog, products, members, and custom pages can be included or excluded from the auto-generated sitemap, and the platform supports submission of multiple sitemaps to Google Search Console. The platform's llms.txt support is the same story as Webflow and Framer — no native support, workaround via redirect rules.

The AEO Migration Playbook by Platform

Most operators who audit their no-code platform's AEO ceiling come away with a question: do we migrate, do we workaround, or do we accept the cap? The answer depends on platform, content volume, and competitive intensity in the AI citation market. The playbook we have run with 14 operators across the four platforms in 2025 and 2026:

1. Audit current schema coverage with Google Rich Results Test and Schema.org Validator. Run 20 sample URLs across page types. Document which schema types are present, which fields are populated, and which are missing. Most no-code sites discover that their actual schema coverage is half of what they assumed it was.

2. Pull current citation share data from a tracking tool (Profound, Otterly, Peec, or Ahrefs AI). Establish the baseline citation rate per page template and per query category. The migration math only works if there is measurable citation upside.

3. Score current platform against the AEO capability matrix. Use the table from earlier in this article. For each capability gap, decide whether the workaround is acceptable or whether the gap is structural.

4. For Squarespace operators with more than 100 pages or measurable AEO competition, plan migration to Webflow or Framer. Webflow if content volume exceeds 200 pages or you publish multiple template types. Framer if content stays under 200 pages and design control matters more than CMS depth.

5. For Webflow operators, deploy template-level JSON-LD as a 30-day project. The lift on Article schema alone has averaged 28 percent citation rate increase in our cohort. Add llms.txt via Cloudflare Worker. Segment sitemaps if content exceeds 1,000 pages.

6. For Framer operators, audit auto-emitted schema against actual page content. Where defaults are insufficient — long-form articles, structured data products, FAQ pages — deploy custom JSON-LD via HTML embed. Plan migration to Webflow if content scales past 200 pages.

7. For Wix Studio operators, invest in Velo skills or hire a Velo specialist. The platform's AEO ceiling is high once Velo is in the mix. Without Velo, Wix Studio sits at roughly Squarespace's level for any non-trivial AEO program.

8. Track post-deployment citation share weekly for 90 days. Citation lift typically appears at 21 to 45 days after deployment, with the steepest gains in days 45 to 75. Earlier signals (Google AI Overview impressions) appear faster than ChatGPT and Perplexity citation rate changes.

Edge Cases: When the Platform Choice Reverses

The default ranking above — Webflow first, Framer second, Wix Studio third, Squarespace last — holds for most B2B and DTC marketing programs. Three edge cases reverse the order.

For a wedding photographer or visual portfolio site under 30 pages where the brand differentiation is visual and AEO citation share is a marginal concern, Squarespace's template ecosystem still beats every alternative on time-to-launch. The AEO ceiling is irrelevant if AI citations are not in the top three demand-gen channels. For a design studio's own marketing site where the build process is the marketing pitch, Framer beats Webflow because the brand signal of "we ship on Framer" matters more than the citation ceiling. For a multi-location service business with 50 to 500 local pages, Wix Studio's combination of Wix Bookings, multi-location address management, and per-location landing pages can beat Webflow because the AEO surface area for local AI search is well-handled by Wix's local schema defaults.

The general principle: if your AEO competition is high and citation share is a measured demand-gen channel, the platform ceiling matters. If AEO is not in the top three demand drivers, platform choice is dominated by other factors — and the no-code platforms are roughly interchangeable on the things that drive those other factors.

What the Platforms Should Ship Next

Looking at the four platforms against the AEO capabilities operators actually need in 2026, the gaps are visible and shippable. Squarespace needs per-template Code Injection on blog and product collections, hand-editable robots.txt with per-user-agent rules, and a llms.txt asset slot at the project root. Webflow needs native llms.txt support and a JSON-LD validator inside the Designer that flags malformed schema before publish. Framer needs a sitemap segmentation feature for multi-template sites and a path to scale past the current CMS ceiling. Wix Studio needs to surface its Velo capabilities more visibly to non-developer users — the platform's AEO story is hidden behind a developer environment that most subscribers never enter.

Platform-vendor incentives are aligned with shipping these features. Sites that get cited more in AI search drive more value to subscribers, more subscriber retention, and more upmarket migration to higher plans. The vendor that ships the cleanest llms.txt support and per-template schema validation in late 2026 will have a real differentiation story against the others. Based on Google Search Central's recent guidance on structured data and the way AI Overviews surface results, the bet on AEO-native platform features is structurally sound.

The Webflow vs Squarespace decision is the one that matters most because it captures roughly 60 percent of the no-code market between the two platforms. For any operator running a content-heavy marketing program, betting on AI citation share as a demand channel, or planning to publish original research that should be cited by LLMs, Webflow is the correct platform choice in 2026. Squarespace remains a strong choice for visual-led small business sites where AEO is not a measured channel and probably will not be for several years.

Takeaway: The platform you picked in 2019 set a ceiling on what your AEO program can achieve in 2026. Webflow's per-template JSON-LD, hand-editable robots.txt, and partial sitemap segmentation make it the default for serious no-code AEO. Framer's strong defaults and clean SSR delivery make it the right choice for marketing sites under 200 pages. Wix Studio closed most of its historical gaps and is competitive once Velo skills are in the mix. Squarespace remains the platform where operator-grade AEO programs hit a hard wall around 100 pages — the per-page Code Injection workflow does not scale, robots.txt is locked to a fixed template, and llms.txt requires an external edge. Audit your platform against the capability matrix, run the migration math against citation share, and ship the JSON-LD stack your platform allows before the ceiling becomes a competitive loss.