Rehearse hard conversations
before you have them.
Firing someone for the first time. Asking your manager for a raise. Telling a peer their work isn't good enough. Practice it out loud, with an AI counterpart who pushes back, then read a coached critique of how it actually went.
A flight simulator for the conversations that matter.
You pick a scenario. You set who you're talking to — supportive, skeptical, adversarial, highly emotional. You set the power dimension — more senior, peer, more junior, external. You write a few lines of context, if you have it.
Then you talk. Out loud. The counterpart talks back, in character, in real time. They push back. They go quiet. They get heated. They do the thing the real person is going to do.
When you're done, you get a critique. What landed. What hurt. What to say instead, in the moment you blinked. No vague encouragement. No therapy speak. The thing a good coach would tell you, if you had one.
Three steps. About ten minutes per rehearsal.
- 01
Pick a scenario.
Choose from a library of real workplace conversations — performance, comp, termination, conflict, boundaries. Or describe your own in a sentence.
- 02
Have the conversation.
Tap to start, then just talk. The counterpart responds in real time, in character. Five to fifteen minutes is typical. You can redo a turn if it ran off the rails.
- 03
Get a coached critique.
A short, direct read of how it went — six dimensions, scored. The moments that landed. The moments that hurt, with a specific phrase to try instead.
The only honest way to build this app.
You will say things in here that you would not say anywhere else. The thing you would tell your manager if you weren't worried about the consequences. The thing you wish you'd said in the meeting last week. We treat that with the seriousness it deserves.
No transcripts. No audio. Ever.
Your rehearsals are not recorded, not stored, and not used to train any model. Voice traffic flows directly between your device and our voice provider. Nothing about a conversation persists once you close the screen.
No accounts. No identity.
There is no email, no password, no profile. Your device gets an anonymous identifier the first time it launches the app, and that identifier is never linked to you.
No retention from our voice provider.
We use OpenAI's Realtime API in a zero-data-retention configuration. Their thirty-day request log is disabled for our organization, contractually.
What we do see.
Anonymous, aggregate counts: how many sessions started, how many completed, which scenario categories. No content. No identity. Used to keep the lights on, not to study you.
First session free. Then nine ninety-nine, once.
You get a complete first rehearsal — the conversation, the critique, all of it — before deciding. If it's useful, unlock the rest with a one-time purchase. If it isn't, you owe us nothing.
- Unlimited rehearsals across all scenarios
- Custom scenarios you describe in your own words
- Coached critiques on every session
- iOS, iPadOS, and macOS — one purchase covers all three
- No subscription. No upsell. No ads.