WoskKnow where
you stand.
You screenshot the conversation. Wosk reads between the lines and says the thing your friends soften.
Tinder · Instagram · iMessage · WhatsApp
Three steps, and about
fifteen seconds.
- 01
Screenshot the conversation.
Up to five images, oldest first. WhatsApp, Messenger, Instagram, iMessage, Tinder — any of them.
- 02
Tap which side is you.
One tap, so Wosk knows whose messages are whose. It never asks who either of you are.
- 03
Read it.
A verdict, how confident it is, and a number out of 100 for how invested he actually sounds.
A verdict, with a
number on it.
Situationship. Breadcrumbing. He’s avoidant. Whatever it is, Wosk names it and tells you how sure it is.
Underneath the verdict is the reasoning: which messages, sent when, and what was never answered.
- Not always bad news. A conversation that is fine gets told it is fine, in green, with the same detail.
- Confidence is shown, not hidden. A thin conversation returns a low-confidence read and says so.
- It knows when to stop. If it finds threats, pressure or coercion it drops the jokes and the score, and shows support numbers instead.
Receipts, not vibes.
His line on one side, what it actually reads like on the other. Every flag points at the message it came from.
“swamped right now, i’ll hit you up”
Message 6
He isn’t planning to. This closes the topic without a fight.
“why do we need a label for everything”
Message 15
He does need one. He needs it to be the one where nothing has been promised.
“fair 😏 sleep well”
Message 4
He heard the no and dropped it. It’s the one thing in here you didn’t have to argue for.
From the example read that ships with the app. Wosk is an entertainment app; analyses are AI-generated and are not advice.
The interest gauge
He replies fast and asks nothing. Nineteen messages across nine days and not one question about your life.
The effort balance
Who is carrying this conversation. It is the one everybody screenshots.
Flags, cited
Red and green, each one named and explained against the messages it came from. Tap a flag, jump to the line.
An attachment read
His pattern and yours, drawn from how the two of you text. A reading, never a diagnosis.
Replies you can send
Three drafts in the tone you pick, built from what he actually said, with the reasoning behind each one.
Follow-up questions
Five per conversation, answered from his messages, for when one read is not enough.
Made to be sent
to the group chat.
One tap turns the verdict into a card worth posting — the label, the confidence, the score. Names are blurred by default, and the default is the right one.
His data stays
out of it.
- The screenshot is processed and deleted in the same request. It is never stored and never seen by a person.
- Nothing about the other person is kept, sold or shared. No advertising SDK, no tracking across apps.
- Phone numbers, email addresses and account numbers are stripped before anything is written down.
- You can delete any single analysis, or the whole account, from inside the app. Both take effect immediately.
The whole of it is in the Privacy Policy, and what the AI does and cannot know is in the AI disclosure.
Your first read
is free.
Free
Upload a conversation and see where you stand.
- One analysis
- The verdict and its confidence
- The interest score
Wosk Pro
The whole read, on as many conversations as you want.
- Every message decoded, line by line
- Every red and green flag, with the receipts
- The interest gauge and the effort balance
- His attachment style, and yours
- Replies you can send, and five follow-up questions per conversation
Weekly with a 3-day free trial, or annual without one. The price in your own currency is shown in the app before you buy. Subscriptions renew until cancelled; cancel any time in the App Store or Google Play.
Stop asking the group chat.
Drop in a screenshot and get a verdict with a number on it — red flags, green flags, and what he actually meant — in about fifteen seconds.