In Instagram DMs, FlickReply suggests replies for you: trigger it, pick a tone — casual, funny, friendly — and it reads the conversation on screen and writes options in that register. In dating apps themselves, Tinder, Bumble and Hinge, reply suggestions are not available; what does work there is polish, which fixes your spelling and grammar and gives you 3 to 4 rewritten versions of the line you were about to send.
Nobody agonises over a work email the way they agonise over a first message. The problem is almost never a lack of things to say — it is rereading one line eleven times, deciding it sounds desperate, deleting it, and typing something worse. A second opinion at that exact moment is worth more than any amount of advice.
Type the version you have, tap polish, and read the alternatives. Seeing the same thought phrased three other ways breaks the loop far faster than staring at it does — and one of them is usually the one you meant.
Pick casual or funny before asking for suggestions and the replies come back in that register rather than sounding like a customer service desk. This is the difference between a reply that lands and one that reads as copied from somewhere.
The polish button is a pause button. Fixing the wording of something written in a bad ten seconds is frequently the point at which you decide not to send it at all.
Real messages, typed the way people actually type them, and the versions that come back.
A first message you have rewritten four times
hey! saw ur into climbing too, ive been going to the place near central for like 6 months now, ur profile made me laugh btw
Nothing about the content changed. Fixing the punctuation and breaking the run-on is the whole difference between this reading as careless and reading as relaxed.
Declining something without sounding blunt
cant do friday, got too much on. maybe next week idk
The original is not rude, it just reads as though you could not be bothered. This is the most common reason people reach for polish.
Worth knowing before you install rather than after.
If the original message is ambiguous, the rewrites will be confidently ambiguous in 3 to 4 different ways. Polish improves how a message reads, not whether the underlying thought was clear.
If you write in a distinctive way on purpose — clipped, blunt, very informal — the rewrites will smooth that out. Some people polish everything; others only use it for work messages for exactly this reason.
Nothing changes until you tap. If you want mistakes fixed silently while you write, that is autocorrect, and FlickReply does not replace it.
FlickReply is an Android input method, so the system hands it every text field on the device — the keyboard itself, the whole tools panel, and polishing a message you have written all work in dating chats, your browser, your notes app and anything you install next month. Only reply drafting is per-app, because it reads the conversation on screen, and that works in WhatsApp and Instagram today.
Partly, and it is worth being exact. Polish and correct works in all of them — write your message, tap polish, and it fixes the spelling and grammar and offers 3 to 4 rewritten versions. Reply suggestions, where FlickReply reads the conversation and proposes what to say, currently work in WhatsApp and Instagram only, so in Instagram DMs but not inside the dating apps themselves.
You choose the register before it writes. Casual and funny produce noticeably lighter results than professional, which is the tone most AI writing defaults to and the reason so much of it reads badly in a personal conversation. There is no dedicated "flirty" preset — casual and funny are the two people use for this.
They can tell if you send the first suggestion unedited, because it will be slightly too neat and will not carry anything only you would say. Everyone can tell. Use the suggestion as a starting point, change a few words to your own, and it reads as you — that small edit is the entire difference and it takes seconds.
Text is only sent when you trigger an AI action: polish sends the message you wrote, a reply suggestion sends the conversation on screen. Nothing is sent in the background and message content is not retained for training or sold. Privacy mode disables both features and history capture for a session, which is the setting to use for a conversation you would rather not have processed at all.
Free to install, 10 AI requests a day, and the keyboard works in every other app on your phone at the same time.
Download free on Google Play (opens in a new tab)