Some conversations should not have to end.
AfterVoice builds a trained AI version of a person, so the people they love can keep talking to them. Their voice. Their memories. The stories only they could tell. Built carefully, while they are still here to shape it.
Most of what makes a person who they are never gets written down. The way they answered the phone. The advice they would have given. The story behind the photo on the wall. When they are gone, all of it goes with them.
AfterVoice is a way to keep some of that accessible. We sit down with someone over a series of conversations, build a careful archive of who they are, and train a private AI model on it. Their family gets a way to keep asking questions long after the original conversations ended.
Who builds one
People facing a diagnosis who want to leave more than photos and a will.
Parents and grandparents whose stories have only ever been told once.
Families who realized, with someone they already lost, that voicemails were not enough.
How it works
The conversations.
Twelve to twenty recorded sessions with a trained interviewer. In person or over video. We cover the big stuff and the small stuff. Childhood, love, work, regrets, opinions, the inside jokes. The point is not to make a documentary. The point is to capture how someone actually thinks and talks.
The archive.
Recordings, letters, photos with the stories behind them, voice notes, anything else worth keeping. Organized, transcribed, and preserved.
The model.
A private AI is fine-tuned on the archive. The voice is cloned from studio-quality recordings. Nothing is shared with third parties. Nothing trains anyone else's model.
The interface.
The family gets a private app. Text or voice. Access is controlled by the person while they are alive, and by whoever they choose afterward.
What it is
A way to keep accessing someone's stories, voice, and way of thinking. Built from what they chose to share. A continuation of a particular kind of conversation.
What it is not
The person. It does not have consciousness. It does not know things they did not tell us. It can be wrong. It is not a substitute for grief, for therapy, or for being present while the person is still here. We will say so plainly, and so should anyone who uses it.
Privacy, built in
- —The archive belongs to the person and whoever they designate.
- —Models are private. No third-party training, no data sharing, no advertising.
- —Self-hosting available for families who want it.
- —A defined shutdown protocol, if the family ever decides to close the archive.
We built this assuming you would read the fine print. So we kept it short.
Pricing
First conversation
Ninety minutes, no charge. We talk about whether this is the right time, and whether we are the right people to do it.
Foundational archive
Twelve sessions, voice training, baseline model. Quoted after the first conversation.
Extended archive
Twenty-plus sessions, deeper personalization, family interface. Quoted after the first conversation.
Annual stewardship
Hosting, updates, family support.
We do not quote on a landing page. The work is too specific. The first conversation tells both of us what this should look like.
A note on grief
We work with grief, not around it. A lot of the families who use AfterVoice also work with a therapist, and we think that is a good idea. This is a tool for memory and connection. It is not a way to skip mourning, and we will tell you honestly if we think it is not the right fit for what you are going through.
Questions
If you have been thinking about this, it is probably worth a conversation.
The first one is on us. No pressure to continue. We will tell you if it is the right time.