Framer has made real progress by introducing localisation features, but in practice it still behaves like a translation tool with a localisation label. On small sites this is fine. On large, multi-region production sites it quickly becomes a bottleneck.

Real localisation is about adapting the full experience per region, not just swapping words. In daily use on a large platform, I repeatedly run into limitations around region-specific content, weak search and filtering, one-directional AI translation, poor long-form reliability, and the inability to edit source content directly while reviewing translations.

These gaps slow teams down and force avoidable workarounds.

The attached PDF contains the full breakdown of the issues I experience with concrete examples, and detailed recommendations for how Framer could evolve from translation-friendly to a truly production-grade localisation platform.