Roast My Chat

Roast My Chat takes one chat screenshot and returns a read of what the other person meant, a suggested reply in the user's voice, and a scored roast of their texting. The image is not stored and the first roast is free.

Roast My Chat is a single purpose web tool built around one input: a screenshot of a conversation. The screenshot goes to a model, which reads who said what and returns three things. First, an interpretation of what the other person likely meant beneath the words, quoted directly from the messages rather than guessed at. Second, a suggested next message written in the user's own tone, ready to copy, with the option of sending nothing when that is the honest answer. Third, a roast of the user's own texting, built from their worst lines in the exchange, plus a rizz score out of 10. The product suits people who overthink a conversation and want a second opinion before replying. That includes dating app threads, early stage texting, and any exchange where a one letter reply has derailed an evening. It is not a general purpose chatbot: there is no back and forth, no memory between sessions, and no ongoing assistant to consult. The whole interaction is one upload, one set of results, roughly ten seconds end to end. In practice the workflow is short. The user drops a screenshot from any messaging app (iMessage, WhatsApp, Instagram and Tinder are named), picks how hard the roast should land across three levels from light tease to full-on savage, and submits. The tool reads the conversation, then answers the question the user brings to it, whether that is whether someone is interested, what to say next, or how to produce a comeback. The reading and the suggested reply are presented as the useful output; the burn and the score are the entertainment layered on top. On privacy, the site states that the image goes straight to the model and is not stored, and that no account is required for a first roast. It also warns users to blur anything sensitive before submitting. The free first roast makes this an easy tool to test once; regular use is where the value question sits, since a single exchange per visit limits it compared with conversational assistants that can hold context across a longer thread.