Most arguments don't start with disagreement. They start with a message that is factually correct and completely unhearable — because the facts arrived before the feeling.
HearMe takes that message and re-delivers it. Your meaning, your points and your boundaries stay exactly as they were. What changes is the order and the language: her feeling comes first, your part in it gets owned, and the actual point lands at the end where it can actually be heard.
It also strips out the language that quietly turns a conversation into a fight — logically, just, objectively, regardless — and turns absolutes like “you always” into what you actually mean: how it made you feel.
The softening slider decides how far it travels: at 0 you keep almost all of your own wording, at 100 it fully re-voices the message while still keeping the meaning intact.
No sign-up, no tracking, no server-side storage of your messages. Your history is saved on your device only. The AI provider is yours to choose — a free keyless one by default, your own API key, or a model running entirely on your own machine.
It keeps your intent and every concrete point you made — the facts, the boundaries, the requests — and rewrites only the delivery so the feeling lands first. It will not invent apologies, guilt-trip you, or claim to know how she feels.
No. HearMe works out of the box on a free, keyless AI provider. You can also paste your own key for OpenCode Zen, Groq, Google AI Studio or OpenRouter, or point it at a local Ollama or LM Studio server.
There is no account and no analytics. Your history lives only in your own browser and can be cleared at any time. Requests go through the app's own proxy, and no API key is ever bundled into the app.
Yes. HearMe is an installable PWA — use your browser's Add to Home Screen option and it launches full-screen with an app icon, and the app shell still opens offline.