Kagi Translate handles Cypriot Greek input well. I tested "Ίντα κάμνεις ρε κουμπάρε, εν θέλω να πάω σινεμά σήμερα, πελλός είσαι;" and got back "How's it going, man? I don't want to go to the movies today, are you crazy?", with a note that it captured the Cypriot dialect using standard colloquial English. The model clearly recognizes Cypriot as distinct from Standard Modern Greek (SMG).
The gap is the other direction. There's no way to translate into Cypriot Greek. The target dropdown only offers Standard Modern Greek, so anyone wanting to produce Cypriot output (writing to family, replying to a Cypriot friend, localizing content, learning the variety) gets SMG instead, which reads as foreign or overly formal to Cypriots.
Claude responded: This is the same asymmetry Kagi already handles for other languages: Portuguese (Brasil) and Portuguese (Angola) both work as source and target.This is the same asymmetry Kagi already handles for other languages: Portuguese (Brasil) and Portuguese (Angola) both work as source and target. Cypriot Greek should work the same way, available as "Greek (Cyprus)" on both sides.
Add "Greek (Cyprus)" as a sibling target option under Greek in the dropdown, on both source and target sides. Detection already works, this just makes the distinction explicit and selectable.
Concrete flows this enables:
- A non-Cypriot writing a WhatsApp message to a Cypriot friend or partner picks Cypriot Greek as target and gets something that sounds native, not textbook
- A learner studying Cypriot can translate English into the variety they're actually trying to learn, instead of having to translate to SMG and then mentally convert
- Content creators or businesses localizing for Cyprus specifically can produce Cypriot output rather than SMG that reads as imported
- Journalists or translators working on Cyprus-focused material can draft in the right register
