Framer Limitations, Honestly: When to Move to Code (2026)

Last updated: 29 October 2026

Framer is excellent for marketing sites your team edits without a developer, but it is not an application platform. It has no built-in accounts for site visitors, no server-side logic of your own, and hard ceilings on pages, CMS items and bandwidth. When a site needs logins, business logic, secret API keys or more than Framer's maximums, it is time to move that part to code.

I build in Framer and in code, so I have no reason to oversell either. Every limit below was checked on Framer's pricing page and documentation on 29 October 2026.

Where Framer is the right tool

It helps to start with what Framer does well, because the limits only matter if you are outside it.

• Marketing sites, landing pages and portfolios designed on a free canvas, with motion built in.
• Editing without a developer: content editors change pages and publish CMS items themselves.
• Hosting, CMS, redirects and staging in one plan, with nothing to update.
• Speed to launch: a custom site can go live in weeks, not months.

If that is your site, most of this post will not apply to you.

Framer's hard limits, by number

From framer.com/pricing, read on 29 October 2026:

Limit	Basic	Pro	Pro maximum with add-ons
Site pages	30	150	700
CMS collections	2	10	40
CMS items	1,000	2,500	40,000
Bandwidth a month	50 GB	100 GB	2 TB
People with edit access	10	10	10
Site redirects	None	Included	2,500, per Framer's plugin reference
Static files (such as .well-known)	None listed	50	50

Enterprise has custom limits and unlimited seats. For most marketing sites these numbers are generous. They start to bite on large blogs, documentation sites, programmatic SEO with thousands of pages, and big teams.

The limits that are not on the pricing page

1. No built-in accounts for site visitors

Framer's member features are about your team: who can view, edit and publish the project. I found no native feature for visitors to sign up, log in or see member-only pages. Those come from marketplace plugins such as FrameAuth, Outseta and Thenty, which add sign-in, gated pages and paid access through services like Stripe. That works for a simple membership or a gated resource library. It is a third-party dependency, though, and not a foundation for an application with user data.

2. No server-side code of your own

Framer pre-renders pages and runs your code components as React in the browser. Code that touches the browser has to wait until the page loads, and anything you put in the front end is visible to visitors. Framer's own guidance for its Fetch feature warns never to expose private credentials in a public URL. If a feature needs a secret API key, a database write or a payment webhook, it needs a backend somewhere else.

3. Code is an escape hatch, by Framer's own description

Framer's developer FAQ puts it plainly: code is "an escape hatch" for adding small dynamic parts to a site, and if a project needs a lot of logic and components, Framer advises building a normal React application instead. That is the clearest line you will get from Framer itself.

4. Code components and the CMS

Framer's help centre says code components and overrides that reached into the CMS through Framer's internals are no longer supported. If a site depends on custom code reading CMS collections, check it against Framer's current APIs before you build more on it.

5. Selling at scale

Framer is not built for a large catalogue. A few products can sell through an embedded checkout or a payment link; a real store with inventory, shipping and tax belongs on a commerce platform.

6. Team workflows

Up to 10 people can have edit access on Basic and Pro. Pro adds staging and branching with previews, which covers most marketing teams. Engineering teams used to code review, automated tests and deploy pipelines will find the workflow different from a code repository.

What you can do before moving to code

• The Server API. Framer's Server API lets a script on your own server update and publish a project without Framer being open, for example to sync an external database into the CMS.
• Fetch. Pull public data from an API into a page without code.
• Plugins and code components for small dynamic parts: calculators, forms, embeds.
• Add-ons. Pages, CMS items and bandwidth scale on Pro up to the maximums above; Advanced hosting, at $200 a month billed yearly, adds multiple sites and custom headers under one domain, with up to 6 rewrites.

Signs it is time to move to code

• People need accounts, and what they see depends on who they are.
• The site stores or changes data: bookings, orders, applications, dashboards.
• Features need secret keys, webhooks or scheduled jobs.
• You are near a hard maximum, such as 700 pages or 40,000 CMS items, and growing.
• The marketing site and the product need to share components, data or sign-in.
• A developer team wants code review, tests and its own deploy pipeline.
• Compliance or procurement requires hosting on your own infrastructure.

One of these is a reason to look closely. Two or more usually means the logic belongs in code.

The setup most teams end up with

Moving to code rarely means abandoning Framer. The common pattern is to keep the marketing site in Framer, where marketing can edit it, and build the product in code on a subdomain, such as app.yourcompany.com. Each part sits on the tool built for it, and the marketing team never waits for a developer to change a headline.

When the marketing site itself needs logic, the whole site moves to a framework such as Next.js, with a headless CMS so editors can still publish content.

What moving from Framer to code involves

1. Inventory every URL on the Framer site, so nothing that earns traffic is lost.
2. Rebuild the pages in the new framework, reusing the design.
3. Move the content into a headless CMS or files.
4. Redirect every changed URL with a permanent redirect, and keep titles and descriptions.
5. Switch the domain once everything is tested, and keep the redirects for at least a year, as Google advises.

The redirect map guide covers step 4 in detail.

What each option costs

Option	Running cost	Build
Stay on Framer	Pro $30 a month billed yearly, plus seats and add-ons	A Framer site on my 14-day sprint, $1,999
Marketing site in Framer, product in code	Framer plan plus hosting for the app	A web app from $1,999 in 21 days, or a SaaS MVP from $2,500 in 30 days
Everything in Next.js	Hosting from a free plan to Vercel Pro at $20 per developer seat a month	A custom-coded website for $1,999 in 14 days

Vercel's Hobby plan is free but only for personal, non-commercial use. All prices checked on 29 October 2026.

My honest take

Most marketing sites never hit Framer's limits, and moving them to code would only make every change slower and more expensive. The ones that should move are the ones that have quietly become products: they hold user data, run business rules or need secrets. For those, Framer's own advice and mine agree: build that part as a proper application, and keep Framer for what it does best.

If you are deciding now, Framer vs Next.js sets the two side by side, and Framer pricing explained covers what each plan includes.