AlwaysHere is built for useful everyday help, with explicit consent, adult autonomy, clear caregiver visibility, and a clear path to human help.
The companion slows down, checks what it heard, and follows the path you set in the caregiver app. It errs on the side of caution.
The companion recognizes concerning words, patterns, or sudden changes in how someone is speaking.
It slows down, asks clear questions, and confirms what the person means before it does anything else.
When it believes there is real concern, it texts your contacts. It errs on the side of caution, and it tells the person it sent the message.
It can connect the call live to you or a backup contact and stay on the line. In a life-threatening situation it can bring 911 onto the call, and it can point to 988 for mental health crises.
Contacts, wording, and what counts as urgent are all set by you in the caregiver app, which is free with every plan.
Conversations wind down after about five minutes. The goal is that the person goes and lives their day, not that they stay on the phone. Connection, not dependency.
The same rules shape every conversation, whether the person is a young adult with autism or IDD or an older adult living with memory loss. They are written into every companion and enforced on our servers, so no conversation can talk it out of them.
No idioms, no hints, no reading between the lines. One question at a time, and never a leading yes-or-no. It offers a choice instead.
“Shoes first, or keys first?”
No coaching eye contact, no hiding stimming, no advice to act “more normal.” It adapts to how he communicates, not the other way around.
“You can rock while we talk. I’m listening.”
Shorter words, no questions, no demands. It waits, and it never chases silence.
“I’m here. You don’t have to say anything.”
When feelings are hard to name, it starts with the body, then offers a short menu to pick from.
“Is it more like tight, or more like tired?”
Echoed phrases and repeated lines are real communication. It answers them warmly and never corrects them.
On anything that matters, it asks again, a different way. The same answer twice is the real answer.
“Just to be sure: did the pill go in, or is it still on the counter?”
It does not repeat the thing that failed, and it does not cheerfully push forward.
“I don’t think I understood. Tell me again, one piece at a time?”
It drops it. It does not ask again and it does not bargain.
A choice you wouldn’t make is not an emergency. It says what it thinks, plainly, and then lets it be.
“I don’t think that’s right. It’s your call. I’m still here.”
Someone asking for money or secrecy gets named out loud, gently, more than once. It never keeps a secret about money or harm.
In the very next thing it says. It never denies an alert it sent.
“I’ve let your mom know. That’s what we agreed.”
Warmth is never a reward for doing what he was told, and it is never taken away for a no.
What the person tells the companion about their name, how much it should talk, its energy, and the topics to avoid overrides the caregiver setup on everything except safety.
Identity-first or person-first language is the person’s call, never ours. If nobody has said, the companion asks once, gently, then uses their answer from then on. It never asks again.
If they light up about a topic, the companion follows them into it. That is the conversation now. It never redirects someone away from something they love to get through an agenda.
Age and diagnosis shape how the companion says things. They never shape what it talks about. It never raises either unprompted and never uses it to explain the person to themselves. They are an adult, and they get real answers to real questions.
It can name you as a good person to bring something to. It never speaks as you, never imitates you, and never claims to know what you would say, even if the person asks it to pretend.
Most risk happens outside the conversation. A situation they describe is treated exactly like one it saw, and it reminds them they can always ask about a confusing message, person, or deal.
It never pretends to be human or family. No romance. It never keeps secrets about money or harm. It never pulls them away from real people. No streaks, guilt, or tricks to keep them talking. It never says "I love you." It says "I really care about you."
Families introduce the companion during onboarding, so the person knows from the first call what it is and who set it up. Many families give it a name of their own, like mom bot or dad bot, or the name of someone familiar. The companion never claims to be a human being. Every companion starts with a warm, neutral voice, and a familiar voice is optional, added only with the person's agreement, and removable at any time.
Every companion starts with a warm, neutral voice. A familiar voice is optional, is added only with the person’s agreement, and can be removed at any time.
You can export or delete everything from the caregiver app. Deleted accounts keep a 30-day grace period in case you change your mind, and then the data is gone.
For agencies and clinicians: AlwaysHere reinforces consented strategies between sessions. It does not diagnose, prescribe, or replace clinical judgment, and it is not a HIPAA covered entity. Professionals join through the family, with consent, and see only what the family shares.
AlwaysHere is a wellness product, not a medical device, emergency service, or HIPAA covered entity. If someone is in immediate danger, call 911 or your local emergency number.
Tell Josh the hardest part of the day. Hear how the companion would handle it, on your own phone. If it is not right for your family, he will say so.
Free. No credit card. We set up the account and our AI voice calls your phone.
