Lifestyle — reply assistant for real conversations
Tavi
For the moment you've typed and deleted the same message four times. Screenshot the conversation, pick a tone, and Tavi hands you three or four replies you'd actually send — then copies one in a tap.
What it does
Features
01
It reads the whole thread
Not just the last line. Tavi follows who said what, where the conversation went, and what is actually being asked of you.
02
Five tones
Funny, Flirty, Bold, Calm or Formal. Switch the tone and get a new set without uploading the screenshot again.
03
Turkish that reads like Turkish
Replies are written in your language, not translated into it. Translated flirting reads wrong, and everyone can tell.
04
It learns your voice
Tavi notices the words and the humour you keep picking, and leans the suggestions your way — so they stop sounding like a chatbot.
05
Not only for dating
A friend you've left on read for three days, an awkward family message, a work follow-up you keep postponing. Formal is a tone for a reason.
06
Bios too
Point it at a profile instead of a chat and get an opener that refers to something actually on the page.
07
One tap to copy
Every suggestion is a single tap to the clipboard, with a haptic to confirm it. Paste, edit if you like, send.
08
History that stays
Past conversations and the replies you were given are kept in your own space, ready to reopen — and gone for good when you delete your account.
Privacy
Your screenshot goes nowhere until you say so — and it is never stored.
Tavi asks permission before the first screenshot ever leaves your phone, and names Anthropic as the AI service that reads it. That permission is checked again on the server for every single analysis, so a "no" is a real no. Nothing that identifies you — no account id, no email, no device id — is ever sent along with the image, and the image itself is held only for the seconds the analysis takes, then dropped. There is no sign-up, and you can withdraw consent or delete everything from Settings.