The state of digital legacy in 2026, drawn from a map that is finally accurate

Digital legacy in 2026 is a real category being described with a broken map. The companies most often named in it have shut down, pivoted, or filed for bankruptcy, and many of the statistics still in circulation are years stale. HereAfter has closed. Eternos became a product for the living. Here is the accurate, dated picture, and what it means for anyone planning ahead while their voice is still their own.

Type "best digital legacy apps 2026" into any search box and read what comes back. Somewhere in the first few results, a confident listicle will recommend a company to record your parent's stories with. Follow the link. A fair number of the time, you will land on a farewell notice, a bankruptcy docket, or a homepage for something that is no longer the thing the article promised. The writing about this industry has not kept pace with the industry, and the gap is not a rounding error. It is the difference between a company you can hand your memories to and a company that no longer exists.

This is not a review roundup, and it is not a prediction. It is a correction. The single most important fact about digital legacy in 2026 is that the public account of it is out of date, and stale information in this particular field is not harmless. People are making decisions about where to store the recordings of people they love, based on a map drawn before several of the landmarks fell down.

Why is every list of digital legacy companies out of date?

Because the category moved faster than the coverage, and it moved in the one direction coverage rarely catches: things ended. A launch is an event with a press release attached. A quiet shutdown, a pivot, a Chapter 11 filing in a courthouse in White Plains, these arrive without a party, and so they are underreported, and so the old article keeps ranking, keeps getting cited, keeps sending a grieving family toward a support address that no longer answers.

There is also a subtler reason, and it is the more important one. This field has always confused two products that share a vocabulary and share almost nothing else. One is memoir capture: you record yourself, or your parent, and you get a book or an archive of stories, made by and largely for the living. The other is simulation: software that talks back in the voice or manner of someone who has died. Both get filed under "digital legacy." They are not the same category, they do not carry the same risk, and a list that blends them tells you nothing you can act on. Half the confusion in the public map is not stale data. It is two different things wearing one name.

A launch gets a press release. A shutdown gets a support email address. The map only ever records the launches.

So the honest starting point for 2026 is humility about the sources. Below is what is actually true right now, each item checked against the primary record this month, with dates attached so you can watch it go stale yourself. Because it will. That is the nature of the thing.

Which digital legacy companies actually died or changed in the last two years?

Start with the closures and the pivots, because these are the entries most likely to still be listed as live.

HereAfter AI, for years the most-cited name in the entire category, has shut down. Its homepage now carries a wind-down notice rather than a product: "It's been an honor to help families record their memories for loved ones, but unfortunately HereAfter is shutting down" (hereafter.ai, fetched 19 July 2026). Families who recorded years of a parent's stories are directed to email support to retrieve them. That is the whole exit. If you read a 2026 guide that still recommends HereAfter as a going concern, you are reading a guide written against a company that is packing up, and it matters enough that we wrote the closure up on its own: HereAfter AI is shutting down, and what to do if a parent's recordings are inside it.

Eternos did not die. It changed what it is. Robert LoCascio's startup, once positioned squarely around preserving a person before death, rebranded to Uare.ai and raised a $10.3 million seed round led by Mayfield and Boldstart Ventures, announced on 11 November 2025 (TechCrunch, 11 November 2025). The reason it gave for the pivot is the single most instructive sentence in this whole essay: it moved toward a "personal AI that sounds like you," for the living, after finding that most of its users were not, in fact, preparing for death. The company followed its actual demand away from legacy. Any 2026 list that files Eternos under "AI to preserve a deceased loved one" is describing a product that walked out of that room over a year ago.

StoryFile, the conversational-video company whose William Shatner demo introduced a lot of people to the idea, filed for Chapter 11 bankruptcy on 5 May 2024 in the Southern District of New York, listing roughly $1.5 million in assets against roughly $10.5 million in liabilities (AI Business, 2024). It did not vanish. It was acquired by Key 7 Investment Company and emerged from Chapter 11 in early 2025, refocused on enterprise use (StoryFile, Key 7 release). The consumer memorial pitch that made it famous is largely not where the company now lives. Again: not dead, but not the thing the old article described.

Notice the shape these three make together. The most-cited name closed. The best-funded name walked away from legacy toward the living. The most-demoed name went bankrupt and came back as something else. This is what a category correction looks like from the inside, and none of it is captured by a list that only knows how to record arrivals.

Three of the category's most famous names, and not one of them is still the company the guides describe.

Some of the newer entrants are best read as experiments rather than institutions, and they are honest about it. Seance AI, built at the software studio AE Studio by designer Jarren Rocks, is explicitly framed for a brief "closure" interaction, described by its own maker as something closer to "an AI-generated Ouija board" than to long-term preservation (Futurism). You, Only Virtual markets a grief companion it calls a "Versona," and describes itself as pre-revenue, running a waitlist and raising through StartEngine (myyov.com). DeepBrain AI's Re;memory recreates a deceased person as an on-screen avatar for a roughly thirty-minute "reunion," delivered in Korea through a partnership with a funeral provider (Voicebot.ai, 19 August 2022). These are real, and they are small, and lumping a thirty-minute funeral-home avatar together with a decade-long preservation promise is exactly the genre error that makes the public map useless. If you are weighing any grief-facing simulation, the questions that actually matter are the ones we lay out in what a griefbot is and is not.

What still works, and what was it ever for?

The steadiest companies in this space are the ones that never promised to bring anyone back. They record the living, and they are quietly the healthiest part of the map.

StoryWorth still does what it has always done: a weekly email prompt to a parent or grandparent, a year of answers, a hardcover book at the end, priced around $99, with the company self-reporting more than a million books printed (welcome.storyworth.com, checked July 2026). Remento captures voice and video and turns it into a book on a similar model, and its founder closed a $300,000 deal with Mark Cuban on Shark Tank that aired in March 2025. Storii runs automated scheduled phone-call interviews for a modest monthly fee, which is a genuinely thoughtful design for an elderly parent who will never open an app. If you are choosing among these, the practical trade-offs are the whole decision, and we walk them in our StoryWorth alternative comparison.

Hold these next to the companion apps, because the public conversation keeps mixing them up. Replika and Character.AI are large, real, and important, and they are not digital legacy products. They are general AI companions. Character.AI spent 2025 restricting open-ended chat for users under eighteen (Axios, 16 July 2025), which is a live safety story, but it is a different story. A companion you talk to for company is not a plan for what outlasts you, and a memoir book is not a living presence. Three different things, one crowded shelf.

So the corrected reading of the vendor landscape in 2026 is not "the category is collapsing." It is more precise and more useful than that. The memoir-capture businesses, the ones with the least dramatic pitch, are stable. The simulation businesses, the ones that promised the most, are the ones that shut down, pivoted, or went through bankruptcy. The boldest claims produced the shortest lifespans. That pattern is the actual state of the industry, and it should tell you something about which promises to trust.

What do the numbers actually say in 2026?

Less than most articles imply, and the honest ones are older than they look.

The most-quoted statistic in the whole field comes from a 2019 Oxford Internet Institute study by Carl Öhman and David Watson, which modeled the future population of the dead on Facebook. At 2018 user levels, they projected that at least 1.4 billion users would die before 2100, and that the dead could come to outnumber the living on the platform within about fifty years (Öhman & Watson, Big Data & Society, 2019). It is a striking number and it is worth citing, but cite it honestly: it is a projection built on a specific scenario, not a measurement, and it is now over half a decade old. Öhman went on to write the closest thing this field has to a definitive book, "The Afterlife of Data" (University of Chicago Press, 2024), which is where a serious reader should start rather than with any single scary figure.

Here is the by-the-numbers picture as of July 2026, dated so you can watch each line age:

• HereAfter AI: shut down, with a wind-down notice live on its homepage and recordings retrievable only by emailing support (hereafter.ai, 19 July 2026).

• Eternos, now Uare.ai: pivoted away from legacy to a personal AI for the living, having raised a $10.3 million seed round announced 11 November 2025 (TechCrunch).

• StoryFile: filed Chapter 11 on 5 May 2024, roughly $1.5 million in assets against roughly $10.5 million in liabilities, since acquired and refocused on enterprise (AI Business).

• Facebook and the dead: at least 1.4 billion users projected to die before 2100, with the dead potentially outnumbering the living around 2070, on 2018 figures (Öhman & Watson, 2019).

• The law: most US states, roughly 46 plus the District of Columbia, have adopted the digital-assets access act, with Louisiana, Oklahoma, and Massachusetts among the holdouts (Uniform Law Commission).

• Germany: six years of litigation, ending 12 July 2018, to establish that a deceased person's Facebook account passes to their heirs (US Library of Congress, 7 September 2018).

Every one of those lines will need editing before long, which is precisely the point. A field where the headline figure is from 2019 and the marquee companies changed shape in 2024 and 2025 is a field that cannot be described from memory. It has to be checked, with a date attached, every time.

What does the law actually say about your digital estate?

The law is the part of the map that moves slowest and matters most, and it too is usually described a decade behind.

In the United States, the operative framework is the Revised Uniform Fiduciary Access to Digital Assets Act, adopted by most states, roughly 46 plus the District of Columbia, with a small group of holdouts including Louisiana, Oklahoma, and Massachusetts (Uniform Law Commission). Its most surprising feature is the hierarchy it sets. At the top is not your will. At the top is whatever "online tool" you configured inside the platform, a Facebook legacy contact, a Google Inactive Account Manager. A checkbox in an app outranks the estate plan your lawyer drafted.

In Germany, the answer arrived through a family's endurance. On 12 July 2018, the Federal Court of Justice ruled that a deceased fifteen-year-old's Facebook account passed to her parents as heirs under the German Civil Code, and that neither post-mortem privacy nor data-protection law nor telecoms secrecy stood in the way (US Library of Congress, 7 September 2018). The court's reasoning was quietly radical: a daughter's messages are a daughter's letters, and letters have passed to grieving families for centuries. That clarity cost the family six years in court.

France went furthest and was noticed least. Its 2016 digital-republic law, at Article 63, gives every person the right to leave binding directives on what happens to their personal data after death, registered with a certified trusted third party (Légifrance). It is close to the right idea. It is also a right that, years on, almost no one exercises.

Four countries, four different answers, and one premise running through all of them: every legal system gives its highest deference to a decision the person made while alive. Where there was a decision, there is an instruction to honor. Where there was none, there is a queue, a courtroom, or a void. We trace this in depth in who owns your digital afterlife, and it is the hinge that connects the legal map to everything a legacy company promises. The law protects what you decided. It cannot protect what you left to chance.

Isn't a shakeout just a healthy market maturing?

This is the strongest objection to everything above, and it deserves to be made at full strength, because it is partly right.

Every young category has a graveyard. Search engines had one, social networks had one, food delivery had one. Companies fail, better ones absorb the lessons, capital finds the survivors, and the field is stronger for the winnowing. On that reading, HereAfter closing and StoryFile restructuring and Eternos pivoting are not a crisis. They are Tuesday. They are what a market does when it is figuring out what people actually want, and the fact that Eternos discovered its users wanted a tool for the living rather than a memorial is not a failure of the category. It is the category learning. A skeptic could fairly say this whole essay is just churn, dressed up as a warning.

Most of that is true. Churn is normal, and a shakeout usually is healthy. But digital legacy has one feature that makes ordinary startup mortality land differently, and it is not a technicality. When a food-delivery app dies, you order dinner from another one. When a company that holds the only recording of your late father's voice dies, and its entire exit plan is a support email address with no promised deadline, an ordinary market event becomes a private catastrophe for a specific family. The asset is irreplaceable. That is the whole difference. The reason to watch this category's failures more closely than others is not that they are more frequent. It is that they are more final, and we set out exactly what to do about that in what happens when a digital legacy company shuts down.

In most markets a shutdown costs you a subscription. In this one it can cost you the last recording of a voice. That asymmetry is the entire point.

So the counterargument holds right up until the moment the asset is a person. After that, "markets mature" stops being a comfort and starts being an argument for a higher standard than any other software gets held to.

What should a company in this category actually promise you?

There is a second industry inside this story, the one that sells preservation rather than a social feed, and I am in it, so this section is written from inside the glass house.

A social network never promised your family forever. Its memorial settings are a courtesy bolted onto a product built for the living. But a whole category of companies exists whose entire pitch is permanence: record your parent, keep the voice, hold the stories for the people who come after. Those companies should be held to a standard far above a social platform, and the last two years show that many of them have not met it. When the most-cited name in the field exits by asking families to email a support address for the recordings they were promised would last, the standard was not high enough.

So here is the standard I believe any serious company in this category owes you, stated plainly enough to be held against us. Everything you put in must be exportable, in ordinary formats, at any time, not as a rescue operation triggered by a shutdown notice. Consent must be documented and specific, given by the living person, never assumed on behalf of anyone who has died. Whatever governs a person's likeness after their death must be written down and binding, so that no one, not the family and not the company, can quietly rewrite who they were. And a company asking families to trust it for generations owes them an honest answer about its own mortality, because that is the one question this category has consistently refused to answer.

That standard is why Afterlife.ai® is built the way it is. Your written data and memories are yours to export as a structured file today, on demand rather than by appeal, and we are candid that extending that same portability to your original photos, audio and voice recordings is a commitment we are still building toward, not a present feature we would dress up as finished. Your Persona is built only from what you chose to put in, with your consent at every step, a living likeness of you as a whole person, never an archive and never an "it." Executor Lock™ exists to answer the governance question directly: at the moment you choose, your Persona is preserved as a faithful snapshot, and no one can alter who you were after that. You can begin with a free build of up to 50 memories, no card, no expiry, precisely because portability and the consent-first design are the promises, not the price.

And here is what a company in this category cannot honestly claim, stated with equal plainness. No one can rebuild a person from the outside after they are gone, and we will not pretend to. A Persona is made by the living person, on purpose, or not at all. We are one company making promises that only decades can verify, which is exactly why the standard above matters more than any feature list, and why the accurate map is worth more to you than the flattering one.

Where does this leave you in 2026?

The state of digital legacy in 2026 is not decline, and it is not hype. It is a real and serious field being navigated with a broken chart. The most-cited names have closed, pivoted, or restructured. The steadiest businesses are the modest ones that only ever promised to help the living record themselves. The headline statistic is from 2019. The clearest law took a bereaved family six years to win. And through all of it runs the one principle every legal system already agrees on: what survives you is what you decided while you were alive.

So the practical takeaway is smaller and harder than any product. Do not choose a company from a list that has not been checked this year. Ask any service you are considering the three questions the closures taught us to ask: can I export everything, right now, in ordinary formats; is my consent documented and specific; and what is your written plan for your own end. If a company cannot answer those, its longevity is a hope, not a promise. If you would rather understand the deeper case before choosing at all, why building a version of yourself is worth doing is where that argument lives.

The map everyone was using is out of date. This one is dated on purpose, so you can tell when it expires. Read it, check it against the primary sources, and then make the one decision no company and no law can make for you, while the making of it is still yours.

Frequently asked questions

Is HereAfter AI still available in 2026?

No. As of July 2026, HereAfter AI has shut down. Its homepage carries a wind-down notice rather than a working product, and families are directed to email support to retrieve their recordings, with no long-term deadline published. If you recorded a parent or grandparent inside it, treat retrieval as time-sensitive. Our guide to HereAfter AI shutting down walks through getting your recordings out, and what happens when a digital legacy company shuts down covers the wider playbook.

What happened to Eternos and StoryFile?

Eternos rebranded to Uare.ai and pivoted away from legacy preservation toward a personal AI for the living, after finding that most of its users were not preparing for death; it announced a $10.3 million seed round on 11 November 2025 (TechCrunch). StoryFile filed for Chapter 11 bankruptcy on 5 May 2024, was acquired by Key 7 Investment Company, and emerged in early 2025 refocused on enterprise rather than its original consumer memorial product. Neither is the company most older articles still describe.

Which digital legacy services are actually reliable right now?

The steadiest businesses in the category are the memoir-capture services that record the living rather than simulating the dead: StoryWorth, Remento, and Storii all remain active in 2026. They produce a book or an archive of a person's own stories, which is a fundamentally lower-risk promise than resurrection. Reliability is less about the brand than the terms, so ask any provider whether you can export everything at any time. Our StoryWorth alternative comparison lays out the trade-offs among the memoir services.

How many people will have died on Facebook, and is that number current?

The widely cited projection comes from a 2019 Oxford Internet Institute study, which estimated that at least 1.4 billion Facebook users would die before 2100 and that the dead could outnumber the living on the platform around 2070, based on 2018 user figures (Öhman & Watson, 2019). Treat it as a dated projection, not a live measurement. It is over half a decade old, which is exactly why any current article should attach the year to it rather than presenting it as today's fact.

Does the law protect my digital accounts after I die?

Partly, and it depends heavily on where you live. In most US states, roughly 46 plus the District of Columbia, a digital-assets access act governs who controls your accounts, and a platform's own online tool can legally outrank your will. Germany's highest court ruled in 2018 that accounts pass to heirs, but that clarity took a family six years to win. France lets you register binding directives in advance. The common thread is that the law protects the decision you made while alive. Who owns your digital afterlife covers this in full.

What should I actually do about my own digital legacy in 2026?

Make the decisions the law and the closures both reward: decisions made in advance, in writing, that you can export and change. Configure the platform tools that outrank your will, write one page of access instructions your family can find, and choose any legacy service on its export policy and consent model rather than its marketing. Then decide the larger question of what of you should outlast your accounts, which is the case made in why building a version of yourself is worth doing.