Fair, transparent, reviewed monthly.

Whichever voice you pick, the companion is the same. The model, the memory, the character — all identical. Voice mode only changes how the companion sounds when it speaks back to you. None of these tiers gives you a “smarter” companion; they’re two voice options on the same conversation.

The Aria voice gives each companion their own voice, designed for that companion and no one else, so Clare always sounds like Clare. It is synthesized on our servers, which is most of what a round costs.

Device uses your phone’s built-in voice instead of one we synthesize on a server, which skips the per-round TTS charge entirely — good for hands-on use, quiet rooms, or stretching credit further.

You can switch modes any time. Credits never expire.

Conversations — apart from Remembrance sessions — aren’t kept after they end. They happen, the companion remembers what mattered, and the transcript itself is gone.

When a conversation holds something worth keeping — a story your children should have, a chapter of your own life told in your voice, a memory you don’t want to lose — switch to Remembrance mode for that session. Remembrances are preserved: return any time to edit, refine, or combine them, share one with someone you choose, or print them as a small book. It costs a little more per round (it covers long-term storage), and it is the only way a conversation persists.

Most of the per-round cost is the voice itself — the speech synthesis that gives your companion her voice, and the transcription that lets her hear you. That is why the two voice modes are priced differently: you are choosing the voice. The AI model that holds the conversation, and the infrastructure that preserves your Remembrances, make up the rest. We review pricing monthly and adjust to track our actual costs.