HeartTold v. HereAfter AI: Read the Guide

01

Retrieval vs Inference: The Most Important Technical Difference

Both HeartTold and HereAfter AI let you ask questions about a loved one. But what happens when you ask a question they never specifically answered is where the two products are fundamentally different — and this difference matters more than any other feature on this page.

HereAfter AI is a retrieval system. When you ask a question, the platform searches across all the recordings the person made and returns the most semantically relevant one. The response comes back in the person's actual voice, playing the original audio. If your loved one recorded a story about their first job, and you ask about their working life, you will hear that story. If they never recorded anything about their experience of losing a parent, and you ask that question, there is no recording to retrieve. The question cannot be answered.

HeartTold's vault is inferential. Across all the conversations captured — everything said in every phone call, every name mentioned, every place and period described — the AI can answer questions that were never asked directly by drawing on what was said. Your grandmother mentioned her father in passing in three separate conversations. She never talked about what he was like as a person. But the AI can piece together a picture from those fragments and answer your question.

This is not a small difference. The questions families most want answered are often the ones that were never explicitly recorded. What was he really like? What did she think about in the years before we were born? What did she feel when her mother died? These are not the questions a person sits down and records answers to. They are the questions a good interview draws out — and the answers live across a hundred small moments, not in a single recording.

02

How Stories Are Captured: Self-Recording vs the Phone Call

HereAfter AI's model requires the loved one to be an active participant in building their own archive. They install the app, work through guided prompts about their life — childhood, relationships, experiences, what they want to pass on — and record their responses directly. The platform then indexes those recordings and makes them conversational for family members later. The content is as rich as the person's willingness to record.

This works well for people who are motivated and comfortable with the process. James Vlahos recorded his father over extended sessions because his father knew he was dying and wanted to be remembered. That context — the knowledge that time is running out and the explicit desire to leave something behind — creates the kind of engagement HereAfter AI's model depends on.

For families where the loved one does not have that same urgency, or simply would not sit down and record themselves into an app, the model breaks down. The vault never gets filled. The conversations the family needed never happen.

HeartTold does not depend on motivation or technical engagement. The phone rings. They answer. A conversation begins. The AI asks questions and follows up: the kind of questions that draw stories out rather than prompting someone to record one. Whether your loved one would sit down with a recording app on their own is irrelevant — HeartTold calls them, and they just have to answer.

03

The Activation Problem: Who Will Actually Use It?

HereAfter AI requires the person building their archive to install the app, navigate its interface, work through its prompts, find moments to record, and sustain that habit over weeks and months. For a highly motivated person — someone who explicitly wants to leave a comprehensive record — this is feasible.

For the people whose stories most urgently need capturing, it is often not. Not because they do not want to share their stories. Because self-directed recording is a particular kind of task that requires a particular kind of person: someone who will sit down with an app and do it consistently, without someone prompting them, without a conversation to carry them through the difficult moments.

User reviews of HereAfter AI also report persistent technical issues — bugs in the recording interface, problems with audio playback, account creation difficulties. For an elderly person navigating the app on their own, these are not minor inconveniences. They are reasons to give up.

HeartTold removes the activation burden entirely. The call comes; the conversation happens; the archive grows. No installation. No prompts to work through alone. No recording setup. The only thing required is answering the phone — which your loved one has done their whole life.

04

“Speaking to the Dead” vs Preserving the Living: The Ethical Dimension

HereAfter AI operates in a category that journalists and ethicists have started calling “grief tech” or “digital afterlife.” The product is explicitly designed to be used after a loved one dies — family members ask questions and hear the person's actual voice responding from recordings. This is powerful and, for many families, deeply meaningful. It is also an area that has attracted thoughtful scrutiny about grief, memory, and what healthy mourning looks like.

HeartTold is oriented differently. The design intent is to capture stories while the person is living — to have the conversation now rather than wishing you had. The queryable vault means that after the person is gone, family members can ask questions and receive answers drawn from what was said. But that is a feature of the archive, not the primary frame of the product. The core use case is: she is still here, and the conversations she has with HeartTold are conversations the family gets to keep.

Neither position is morally superior. Families grieve in different ways, and for many people who have used HereAfter AI, hearing their parent's actual voice answering a question — in recordings made before they died — is genuinely healing. What matters is knowing which frame you are buying into, and whether it is the right one for your family.

05

The Coverage Gap: What Happens When the Recording Does Not Exist

The hardest truth about any retrieval-based system is that it can only return what was put in. HereAfter AI's Starter plan limits users to 20 stories. The Storyteller plan allows 50. Even on the Unlimited plan, the vault contains only what the person chose to record — which is constrained by the hours they sat with the app, the prompts that happened to match their significant memories, and the comfort they felt recording themselves.

A person who lived for 80 years has tens of thousands of memories. The ones they did not think to record — the ones that never came up in 50 prompted recordings — are gone. When a grandchild asks a question 20 years after the person died, and the answer to that specific question was never recorded, HereAfter AI has nothing to retrieve.

HeartTold's phone call format is specifically designed to surface what would not have been recorded: the stories that come out in conversation, not in response to a prompt. The dynamic follow-up questions, the unexpected tangents, the moments when a question about one thing leads to a story about something entirely different — these are the interviews that generate the kind of material a retrieval system can work with. And because HeartTold's vault is inferential rather than retrieval-based, even a partial mention of something can form part of an answer.

06

The Voice: What HereAfter AI Does That HeartTold Does Not

HereAfter AI's most powerful feature is also its most emotionally distinctive: the response comes back in the person's actual voice. Not an approximation. Not a transcript. The recording they made, played back in response to your question. For many families, this is irreplaceable — the specific timbre, the accent, the particular way they paused before something important.

HeartTold preserves voice recordings of the conversations, and the AI can quote from them. But HeartTold's vault does not respond in the person's actual voice the way HereAfter AI does. The response to a question is drawn from the captured conversations, but it is text and quotation rather than audio playback of the original recording.

For families where hearing the actual voice is the primary need, HereAfter AI delivers something HeartTold does not. If the priority is breadth and depth of stories — capturing a lifetime of material from someone who would never record themselves — HeartTold is the more effective tool. The two products are strong in different dimensions.

07

HereAfter AI vs HeartTold: Pricing Compared

HereAfter AI charges monthly, with no annual billing option. The Starter plan is $3.99 per month for 20 stories and 20 photos. Storyteller is $5.99 per month for 50 stories. Unlimited is $7.99 per month with no story cap and MP3 download access. At monthly rates, the comparable Unlimited plan costs $95.88 per year. There is a 14-day free trial on all plans.

HeartTold's Pro plan is $84 per year ($9.99 per month, or effectively $7 per month billed annually), or $79 as a one-time gift for a full year. At the annual billing rate, HeartTold is cheaper than HereAfter AI's Unlimited plan ($84 vs $95.88 per year). HeartTold also includes capabilities HereAfter AI does not: AI outbound calling, PDF memoir export, family tree, and occasions-based triggers.

For families who want the monthly flexibility and are primarily interested in building a voice archive rather than AI phone calls, HereAfter AI's monthly entry point ($3.99) is lower. The right comparison depends on which features actually matter to you.

08

When to Choose HereAfter AI and When to Choose HeartTold

Choose HereAfter AI if the most important thing to you is hearing your loved one's actual voice responding to a question — and if your loved one is willing and able to record themselves into the app. The Dadbot story is real and the product that came from it has genuine emotional power. For a motivated person who will put in the time to record their stories, HereAfter AI can produce something profoundly moving.

Choose HeartTold if your loved one will never sit down and record themselves — but would pick up the phone if someone called. If the concern is capturing as many stories as possible from someone who has never talked about their life in a systematic way, HeartTold is more likely to produce those conversations. And when the questions start coming — from grandchildren who were not yet born when she was talking, from family members asking about things she touched on but never fully explained — HeartTold's inferential vault can reach beyond what was explicitly said.

For some families, both products serve a purpose. HereAfter AI for the parent who will record themselves. HeartTold for the parent who will not, but who will always pick up the phone.