Your Voice Technology
Key Facts
Your KindredMind voice presence is built from three short interview sessions with our setup specialist Sarah during onboarding, about 7 minutes each. Your Instant voice presence is ready after Session 1, and it deepens into your Professional voice presence after Sessions 2 and 3. You can optionally add more Advanced Training sessions later from your dashboard. Real conversations between the companion and your loved one are never used for training.
Your KindredMind voice presence genuinely sounds like you, built from short interview sessions, deepened on your terms. Not a robot voice. Yours.
Start your voice setupWhen a person with dementia calls a family caregiver, they're not calling for information. They're calling for the specific, irreplaceable comfort of a familiar voice. The sound of you, not a nurse, not a care aide, not a generic assistant, you. The voice they've known for 50 years. The voice that has always meant safety.
Generic voice AI doesn't work for this. The moment your loved one hears a voice that doesn't sound like you, the reassurance collapses. The call ends without the comfort it was seeking. And twenty minutes later, the phone rings again.
This is why voice cloning is not a nice-to-have feature in KindredMind. It is the entire thing. Without it, we're just another AI phone system. With it, we're something genuinely new: a way for your voice to be present even when you can't be.
To understand more about why dementia patients are so focused on reaching their family caregiver, our piece on dementia separation anxiety explains the neuroscience behind the repetitive calling.
Your voice presence is built in two clear phases. Your onboarding interview sessions with Sarah build the model. Optional Advanced Training sessions from the dashboard let you deepen it whenever you want.
During onboarding, our setup specialist Sarah holds a short, natural interview with you. You answer easy questions in your own words. Sarah listens and gently follows up. The session captures your warmth, cadence, and natural pauses. It becomes your initial Instant voice presence.
As soon as Session 1 finishes, your voice model is ready. The next call to your loved one's KindredMind number is answered in your voice. The initial model captures the fundamental qualities of how you sound and is already meaningful on a phone call.
Your remaining onboarding sessions with Sarah, each about 7 minutes, give the model more of your voice and more of your stories. After Session 3, the model is retrained as your Professional voice presence. Real conversations between the companion and your loved one are never used at any point.
From your dashboard, the Voice tab lets you add more Sarah sessions whenever you want, to deepen the voice and grow the knowledge base. You can preview your current voice at any time to hear exactly what your loved one hears.
Without a voice presence
With your KindredMind voice presence
Does it actually sound like me, or does it sound like an AI?
The initial Initial Voice is ready after your first short interview session with Sarah, recognizably your voice, warm and familiar, working from day one. After Sessions 2 and 3 during onboarding, the model is retrained as your Professional voice presence. You can optionally add more Advanced Training sessions later from your dashboard to deepen it further. The voice is built only from your interview sessions. Real conversations between the companion and your loved one are never used for training.
What if I don't want to record? Can I skip the voice setup?
You can skip the voice recording during onboarding and add it later. KindredMind will use a warm, neutral voice in the interim. We do strongly recommend completing the voice setup, it's the feature families consistently say makes the biggest difference. You can return to it from your dashboard at any time.
Is my voice data private? Who can access it?
Your voice recordings and model are stored securely and encrypted with AES-256, the same standard used by banks and healthcare providers. We do not sell your voice data or use it to train any general AI models. Call transcripts generated during your calls are also encrypted at rest and deleted after 90 days. We never ask for your loved one's full legal name or date of birth. The only voice audio we ever process for training is yours, captured during your short interview sessions, and real conversations between the companion and your loved one are never recorded for training. You can delete your voice model and all associated recordings at any time from your account settings. For full details, read our privacy and security page.
How does the model get better over time?
Your model deepens whenever you record another interview session with Sarah. Sessions 1, 2, and 3 happen during onboarding. After that, you can open the Voice tab in your dashboard and add another short session whenever you want, to capture more of your range, your stories, or how your voice has changed. Real conversations between the companion and your loved one are never used. Your dashboard shows your current voice status and lets you preview it at any time.
Can I update my voice model if my voice changes?
Yes. You can record a fresh interview session with Sarah at any time from the Voice tab to refresh your voice presence. If you'd like to reset and start completely fresh, that option is available in settings.
Is voice cloning ethical in this context?
This is a question we take seriously. The caregiver creates and owns their own voice model. It is used exclusively for calls from their loved one, authorized by the person who holds legal decision-making authority for their care. In most families this is a spouse, adult child, or the holder of a Power of Attorney for Personal Care — the same person who makes decisions about their loved one's facility, medications, and care approaches. The goal is to provide comfort to a vulnerable person, not to deceive for any harmful purpose. We address the full ethical question honestly on our ethics page.
Can more than one family member build their own voice presence?
Yes, the Care Circle plan is designed exactly for this. Up to four caregivers can each record their own voice and build their own voice presence. Each person gets a separate dedicated number and their own dashboard. When your loved one calls your number they hear you; when they call your sibling's number they hear your sibling. Every voice is completely independent, recorded, stored, and owned by that caregiver. One subscription, shared by the whole care circle. See the Care Circle plan.
The Your Voice model does not have to stop improving when onboarding ends. The Voice tab in your dashboard is one of the things families most often mention when they check back in weeks later.
Your initial Instant voice presence is built from your first short interview session with Sarah, about 7 minutes of natural conversation. It captures the fundamental qualities of your voice: your warmth, your cadence, the emotional tone that makes you sound like you. For many families, the result is already remarkably close from day one.
Sessions 2 and 3 during onboarding give the model more of your voice and more of your stories. After Session 3, the model is retrained as your Professional voice presence. Throughout this whole process, the only audio used is yours, captured during your sessions with Sarah. Real conversations between the companion and your loved one are never recorded for training.
That separation matters for two reasons. First, it keeps your loved one's calls completely private. Second, it means the model is learning only from you, in a calm setting where you are speaking your best, not from a worried call where your voice may not be at its most natural.
From the dashboard, the Voice tab lets you add more Sarah sessions whenever you want. Each session deepens the voice and grows the knowledge base, so the calls in month two can sound a little more like you than the calls in week one, on your terms.
Your voice data is handled with the same care as everything else KindredMind processes. The raw audio captured during your sessions is deleted within 24 hours of your model being ready. Your voice data does not train any general AI model, it is not sold or shared, and it is permanently deleted when your account closes. See our privacy commitments for full details.
The goal of all of it is simple: that every call your loved one makes sounds genuinely like the person they are calling. Not an approximation. Not a robot. You.
Three short interview sessions with Sarah, about 7 minutes each. Your voice model is ready after Session 1. Your loved one's next call is answered in your voice.
Start your setup, it begins in minutes