Framer's animations — appear effects, scroll reveals, hover transitions, variant switches — aren't baked into the HTML. They're produced at runtime by Framer's own bundled JavaScript engine reading instructions embedded in the page. That single fact determines everything about what happens to your animations when you export, because export tools make a choice about that runtime, and the choice is the whole story.
Option A: keep the runtime — animations are identical
If a conversion keeps Framer's runtime intact (the "pure" or fidelity-preserving approach), your animations don't change at all, because the exact same engine is still running them. Every easing curve, every stagger, every scroll-linked effect behaves byte-for-byte like the original. The tradeoff is that you're also keeping that runtime's weight — which is why this mode preserves fidelity but doesn't improve performance. We measured that tradeoff across 10 real templates here.
Option B: strip the runtime — animations get rebuilt
If a conversion strips Framer's runtime to make the page lighter and faster (the "hybrid" approach), the animations that depended on that runtime can no longer play as-is. A good converter rebuilds the common ones — appear-on-load and scroll-into-view reveals — using a lightweight CSS-plus-IntersectionObserver layer that reproduces the effect without the heavy engine. This is why hybrid mode can be dramatically faster: it's not shipping the animation runtime at all, just a tiny replacement for the effects that matter most.
What translates cleanly, and what doesn't
The reveal-style animations — fade-up, fade-in, slide-in on scroll — rebuild faithfully, because they map naturally onto CSS transitions triggered by an intersection observer. What doesn't translate 1:1 are the effects that are genuinely runtime-driven in a way CSS can't cheaply replicate: complex physics-based motion, tightly scroll-synced parallax sequences, and elaborate multi-state interactive components. On an animation-heavy site, those are where you'd see a difference in hybrid mode.
How to decide
If your site's animations are the point — an interactive showcase, a motion-design portfolio — keep the runtime and accept the weight (Option A). If your animations are mostly tasteful reveals and your priority is load speed, strip the runtime and let them rebuild (Option B); the difference is usually invisible and the speed gain is real. The safe move is to try hybrid first, compare it against the original visually and on Lighthouse, and only keep the full runtime if you actually see an effect you can't live without losing.