For the “what are we” stage
Upload the chat. Get a verdict.
Reads: who initiates · whether plans land.
Case intake
Which one is you
One read, then gone. Your chat travels over an encrypted connection, the AI reads it once to write your verdict, and it is never written anywhere: no database, no copies, no logs of the content, no account attached to it. Screenshots lose their metadata (location, device, time) on your device before anything is sent.
Nothing is shared unless you create a share link yourself, and that link carries the verdict and the counting stats only: no messages, no names, nothing you pasted.
The finding
01 · Findings
04 · Working in your favor
05 · The next move
Shared case file
Inconclusive
What would help
People who can help
If you are in immediate danger, call 911.
Lab error
You upload the chat, as screenshots or pasted text, and say which name is you. An AI reads the whole timeline the way a brutally honest friend would: who texts first, how long replies take, who makes plans and who dodges them. Then it writes the verdict.
No. The chat goes to the model once and is never written anywhere: no copies, no logs, no analytics on the content. Screenshots lose their metadata on your device before anything is sent.
Everything travels over an encrypted connection, the same HTTPS a bank login uses. There is no account, no database, and no stored copy of your conversation: once the verdict is written, the chat is gone from the pipeline. We never see who uploaded what, and there is nothing on our side to leak, subpoena, or get breached.
The verdict and the counting stats (message totals, reply times, first moves), and only if you create a share link yourself. No messages, no names, nothing you pasted. You can send it around without sending the receipts.
It reads one chat, and a chat is not a relationship. Treat the verdict as a sharp read of your texting pattern, not a ruling on your life. If something real is on the line, talk to the person, not the robot.
