Being a Framer partner would make it easy to answer every stack question with Framer. That would also be bad advice. The honest answer depends on three variables: who edits the site after launch, how much custom logic it needs, and how fast you need to move.
When Framer wins
Marketing sites, launch pages, event sites. Your team edits copy without a deploy pipeline, animations ship in days not weeks, and hosting, CDN and forms come standard. For 80 percent of startup websites this is simply the right tool, which is why we invested in the partnership.
When code wins
Product dashboards, anything with wallets or auth, sites where SEO demands full control over rendering and structured data. Framer is a website tool, not an application framework. Pretending otherwise creates expensive rebuilds a year later.
The questions we actually ask
Who edits the site next Tuesday, and do they write code? Does anything need auth, a wallet, or a database? Does the SEO strategy depend on structured data and rendering control? Is the site expected to become the product later? Four questions, answered honestly, settle the stack in ten minutes. Taste never enters it.
Price it over two years, not two weeks
Framer is cheap at launch and stays cheap while the site stays a site. The cost appears the day product features creep in: custom logic bolted onto a website tool, workarounds stacking until a rebuild. Custom code is the mirror image, more expensive on day one and nearly free to extend. Choosing the stack is choosing which curve you would rather be on.
Migrating later is not a failure
Some of the best-run projects we have seen launched on Framer and moved their product surface to code a year later, on schedule, without drama. That is the system working. A shared design system makes the move boring: same tokens, same components, different renderer. Plan for the possibility on day one and the migration becomes a line item, not a crisis.
The hybrid we ship most
Framer for the marketing surface, code for the product, one design system across both. Users never notice the seam, teams get the editing speed where it matters, and engineers stop being a bottleneck for copy changes.