When to move your marketing site off Next.js (and when not to)
Move your marketing site from Next.js to Framer when marketers wait on engineers for every page and copy change, and the site is a few hundred pages with a CMS under Framer's 40,000-item ceiling. Stay on Next.js when marketing pages read product data or auth, need per-visitor personalization, run as a very large programmatic page set, or your engineers already ship pages fast. A Desses migration starts at $8,000 / €7.000, and a typical 30 to 75 page site runs €12.000 to €25.000.
Category
Platform
Reading time
8 min

Move your marketing site off Next.js when the people who write the pages can't publish them without an engineer, and the site is a few hundred pages with a blog that fits inside Framer's CMS limits. Stay on Next.js when the marketing pages read product data or user sessions, personalize per visitor, or generate tens of thousands of programmatic pages. Hosting cost barely enters into it. The cost that decides it is the queue between marketing and engineering.
I build both. I'm a Framer Partner, so a Framer migration is revenue for me. I also build marketing sites in Next.js and Astro, with Sanity as the CMS, when Framer is the wrong tool, and apps like ChargeBlock's CRM and energy platform on Next.js, Convex and Clerk. Both answers pay the same, so the one you get depends on your site.
Every Framer price and limit below was read off framer.com/pricing on 18 November 2026. Framer changes plans often, so check before you budget.
Is Framer or Next.js better for a marketing site?
For a site that marketers edit weekly, Framer. For a site that behaves like part of the product, Next.js.
Framer | Next.js (on Vercel) | |
|---|---|---|
Who ships a copy change | The marketer, in the editor | An engineer, through a pull request and deploy |
Rendering | Every page pre-rendered to HTML (Framer Help) | A per-route choice your team makes |
CMS | Built in. Pro: 10 collections, 2,500 items, buys up to 40 and 40,000 | Whatever headless CMS you wire in, no platform ceiling |
Site pages | Pro buys extra pages in blocks of 100 ($20 each), up to 700 | No ceiling |
Custom code | React 18 code components inside the canvas (Framer Developers) | Anything |
Auth, product data, per-user content | No | Yes |
Platform bill | Pro $30 a month, extra editors $20, content editors $10 | Vercel Pro $20 per developer seat plus usage (Vercel) |
Markdown for AI agents | Every optimized page, via | You build the route |
The platform bills are close and small. A Next.js marketing site on Vercel is cheap to host. What it costs is the engineer time behind every edit, which nobody puts on an invoice.
Should your marketing site be separate from your app?
Usually yes, once marketing ships more than a page a month. A marketing site and an app change for different reasons, at different speeds, by different people. Keeping them in one repo ties a homepage headline to the app's release cycle, its code review and its test suite.
The pattern I see in B2B and SaaS teams goes like this. The marketing site started in the product repo because an engineer built it in a weekend. Two years later, marketing has a backlog of landing pages, and every one is a ticket that competes with product work. Product work wins, as it should. So the landing pages don't ship.
You can split them and keep one domain. Framer's Multi Site rewrites can serve an external HTTPS app or docs site under a path on your domain, with the URL unchanged in the browser. That sits in the Advanced Hosting add-on, $200 a month with up to 6 rewrites. More often the app already lives on app.yourdomain.com, and the marketing site moves to Framer on the root without touching it.
When should you move from Next.js to Framer?
When three or more of these are true:
Marketing waits on engineering for copy changes, new landing pages or blog posts.
The site is mostly static pages plus a blog, case studies and maybe a changelog.
No marketing page needs a logged-in user, live product data or per-visitor content.
The CMS holds hundreds or a few thousand items, well under 40,000.
A designer owns how the site looks, and every design change today goes through someone else's hands.
Nobody on the engineering team wants to own the marketing site.
That last one is the quiet signal. When the site has no owner in engineering, it decays: dependencies drift, the build breaks on a Friday, and the blog template still has last year's nav.
Watermelon (watermelon.ai) went from Next.js to Framer with me and has been live since late September 2026.
When should you stay on Next.js?
Stay if any one of these is true. Each is a real reason, and I turn down the migration when I hear one.
Marketing pages read product data or auth. Pricing pulled from your billing system, a signed-in header, a dashboard preview with the visitor's own numbers. Framer pages have no access to your sessions or your database.
Heavy personalization. Content that changes per account, per segment or per experiment beyond simple A/B tests. Framer's Convert add-on runs A/B tests, which covers most marketing needs and nothing past that.
Very large programmatic page sets. Framer's CMS has a hard ceiling of 40 collections and 40,000 items. Integration directories, location pages or template galleries that run past that belong in code.
Your engineers already ship pages fast. If a landing page goes live the same day and marketing is happy, a migration buys you risk and a new tool to learn.
You need full control of the request. Custom headers per route, edge middleware, server actions, your own caching. Framer's Advanced Hosting covers a handful of headers and rewrites, and the rest is closed.
Staying in code doesn't mean staying with the site you have. If the marketing site needs a rebuild and still belongs in code, that's a custom website in Next.js or Astro, with Sanity as the CMS your team edits. If the app needs work, that's AI and custom development, the same stack I used for ChargeBlock.
What does a Next.js to Framer migration cost?
Two numbers: the Framer plan and the rebuild. The plan is small. The rebuild is the project.
Cost | |
|---|---|
Framer Pro | $30 a month |
Extra editor / content editor | $20 / $10 per month each |
Locales | $20 per locale per month, up to 20 |
Advanced Hosting (rewrites, page-by-page move) | $200/mo |
Desses migration project | from $8,000 / €7.000 |
Typical 30 to 75 page site, design and build | €12.000 to €25.000 |
Ongoing design and build after launch (Growth retainer) | $4,500 / €4.000 a month |
Nothing converts Next.js into Framer. React pages don't import as canvas layouts, so every page gets rebuilt, which is also the moment to fix the design debt the old site carried. Most of my migrations include a redesign for that reason. The website design and the Framer build run as one engagement.
What moves the price inside that band: page count, how many distinct templates the site has, how many CMS collections, the number of languages, and how much custom interactive code has to come across as code components. Every migration ends with recorded tutorials at handover and 30 days of support after launch.
What does the move involve?
The platform switch is the last step. Most of the work is making sure nothing the old site earned gets dropped.
URLs and redirects
I pull every live URL from the sitemap, Search Console and the Next.js routes, then keep paths identical where I can. Anything that changes gets a permanent redirect: Framer sends a 308, which Google treats the same as a 301. Framer's redirects support wildcards and capture groups, and Framer's own docs warn that renaming a path in the canvas or CMS doesn't update existing redirect rules. Every rename needs a rule beside it.
CMS content
MDX files, a headless CMS or JSON in the repo all have to be mapped to Framer collections. Rich text, slugs, dates, authors and references get checked after import. Embedded React components inside MDX don't survive and need a plan per component.
SEO parity
Titles, meta descriptions, canonicals, Open Graph images and JSON-LD carried over page by page, then a crawl of the staging site against the old one before DNS moves. I wrote up the AI-citation side of this in redesigning without losing your AI citations.
Analytics
Framer has built-in analytics, and GA4 or another tag goes in through custom code. I test page views across client-side navigation, because a pasted snippet can record the landing page and miss every route after it.
Forms
Framer's native forms post to email, Google Sheets or a custom webhook (Framer Help). HubSpot forms embed with a component. Every form gets tested end to end into the CRM before launch.
Custom components
Calculators, interactive demos and pricing toggles move as React code components. They have to be React 18 compatible and self-contained. Anything that called your API from the server needs a public endpoint or gets redesigned.
For a large site there's a slower route. Framer's page-by-page migration serves finished Framer pages and falls back to the old Next.js site for every path that hasn't moved yet.
What do you give up?
Code-level control, mainly. No pull requests on the marketing site, no tests, no type checks, no git history of copy changes. Framer has version history, but the review step your engineers had is gone, and a marketer can publish a typo to production.
You also take on Framer's limits: the 40,000-item CMS ceiling, a closed hosting layer, and a Server API that speaks WebSockets and is still in open beta. If your content pipeline ran through git, it now runs through Framer.
And you take on a vendor. If Framer's pricing or direction changes, you rebuild again. I think that trade is right for most marketing sites. I said the same about Webflow in Framer vs Webflow, and the rendering side is in Framer, Webflow or Next.js for AI search.
Who this is for
B2B and SaaS teams whose marketing site lives in Next.js, often next to the product, and whose marketers wait on engineers to ship a page. Usually a redesign and Framer build in one go, with the app left exactly where it is.
Moving from WordPress or Webflow instead? Those have their own pages: WordPress to Framer and Webflow to Framer.
Who shouldn't hire me: teams whose marketing pages are part of the product, anyone wanting a site built in WordPress or Webflow, a one-pager, or a project under €7.000. If you're weighing the move, send me your URL and a rough page count and I'll tell you whether it's worth doing. Sometimes it isn't.
Sources
Framer, Pricing, https://www.framer.com/pricing (accessed 18 Nov 2026)
Framer Help, Make your site readable by AI agents, https://www.framer.com/help/articles/make-site-readable-by-ai-agents/ (accessed 18 Nov 2026)
Framer Help, How to setup redirects to maintain SEO ranking, https://www.framer.com/help/articles/how-to-setup-redirects-to-maintain-seo-ranking/ (accessed 18 Nov 2026)
Framer Help, Migrating to Framer page by page, https://www.framer.com/help/articles/how-to-migrate-a-site-to-framer-page-by-page/ (accessed 18 Nov 2026)
Framer Help, How to use Multi Site with rewrite rules, https://www.framer.com/help/articles/how-to-use-multi-site-with-rewrite-rules/ (accessed 18 Nov 2026)
Framer Help, Adding HubSpot forms to your site, https://www.framer.com/help/articles/how-to-add-hubspot-forms-to-your-website/ (accessed 18 Nov 2026)
Framer Developers, Code components introduction, https://www.framer.com/developers/components-introduction (accessed 18 Nov 2026)
Framer, Server API update (published 12 Feb 2026), https://www.framer.com/updates/server-api (accessed 18 Nov 2026)
Vercel, Pricing, https://vercel.com/pricing (accessed 18 Nov 2026)



















