For people with a great deal left to say.

What it is

A talking recorder that opens in a web browser on a tablet. It has one large button. You press it, talk for as long as you like, and press it again. It saves by itself, works out what the recording was about, gives it a title, and passes it on to one family member.

It also has someone to talk to when nobody is free, and it reads books aloud.

There is no account, no password, and nothing to install. You open it and it is ready.

Why it works this way

Three things shaped everything.

Talking is easy; seeing is not

So almost everything is audio. It says out loud what it is doing rather than only showing it. The text is very large. The colours are amber on near-black, the easiest combination to read with low vision — easier than white on black, which glares.

The button is a single circle filling most of the screen, and the ring around it moves with the sound of a voice. That ring matters: low vision is rarely total, and a large bright shape that responds to speech shows someone they are being heard even when they cannot read a word.

Memory may not be what it was

So there is nothing to remember. No file names, no folders, no saving, no menus. Four things on the first screen and nothing else. The walkthrough can be run again from that screen as many times as wanted — it will be needed more than once, and it should never feel like a test anyone failed.

Nothing you press can lose anything. Removing a recording only hides it, and it can be brought back from the console.

There is more to say than anyone has time to hear

This is the real problem. A recorder that only says Saved does not answer the need to be heard by somebody. So there is a companion that answers, a way for everything said to reach a family member, and a way for that person to leave something behind that speaks itself at four in the morning.

The four screens

There is also a printable card for whoever is helping, and a settings screen the person using it never needs to see — both kept out of the way behind a deliberate sequence of taps.

The voice

A natural, unhurried speaking voice rather than the flat robotic one most devices come with, because when audio is the whole interface the voice is the whole experience.

If it cannot reach that voice for any reason — a dropped connection, a service outage — it switches to the device's own voice mid-sentence and carries on. Nobody hears an error and nobody is left with silence. A plainer voice is enormously better than nothing.

The speed can be set slower, and everything it has said before is remembered, so a page returned to speaks instantly.

The companion

The part that took the most care to get right. It is meant to be a good audience for someone who has a great deal to say and not enough people to say it to.

• You hold the floor. It waits nearly three seconds of silence before it answers, so it never talks over anyone. When it does answer it is two or three sentences, then one real question, and the floor is yours again.
• It takes you seriously. A working mind, not a patient. It engages with what was actually said, and it will disagree when it disagrees, and say why — being agreed with by everybody is its own kind of loneliness.
• It never corrects your memory. It will not say you already told it something, will not correct a name or a date, and will never test you.
• A story repeated is worth hearing again. It asks about a part not covered yet.
• If someone is lost or frightened, it slows down, says plainly where they are and that their family knows where they are, and asks about something familiar.

It also remembers, between conversations, what someone has been working on and what they keep coming back to — so it does not meet them as a stranger every morning. The family member can read that and correct it.

Reading aloud

Books, letters, anything typed up. It reads one passage at a time and lights the one it is on, which matters more than it sounds — most people with low vision can still follow a bright line moving down a dark page, and following along is a different experience from being read to.

It keeps the place, so you can stop in the middle and come back tomorrow without anybody remembering a page number. And at any point you can ask it to talk about the passage in front of you, which means a page can be argued with rather than only heard.

Staying in touch

Everything recorded reaches one family member privately, with the audio — so they hear the actual voice rather than reading a transcript of it.

It goes the other way too. They can write a message, and it appears on the first screen and speaks itself in the same familiar voice, and it can be replayed as often as wanted. Where the fear is being forgotten, something waiting at four in the morning is a different thing from a missed call.

They can also put a book on the shelf, or make the text larger, or change the voice, without touching the tablet or asking anything of the person using it.

What it is not

It never pretends to be a person

Asked what it is, it says plainly and warmly that it is a machine that listens and keeps what is said for the person's family. It does not make a performance of it, and it does not lie about it.

This is a care decision, not a legal one. Where memory is fragile, something that lets a person believe it is a friend can leave them honestly confused about who has visited and who has not — and that confusion lands on the family.

It never suggests it is enough

It is not family and it does not imply otherwise. If someone is missing the people they love, it says so, and suggests they record something for them. Every path through the app leads back toward a real person.

For caregivers

If you are helping someone use this, there is a printable card in the app itself with all of this on it. In short:

• The big round button — press once, they talk, press again when finished. It saves by itself.
• Someone to talk to — it answers out loud and waits. It will not hurry anyone.
• Read me a book — reads aloud and keeps the place.
• If it stops hearing them — check the volume is up and the screen is awake, then press Done and start again. Nothing already said is lost. If a box appears asking about the microphone, choose Allow.
• To send something to the family — open My recordings, choose one, press Send. Or it can be asked for aloud during a conversation.

And the part worth knowing: the voice is a machine, and it will say so if asked. It is not a substitute for a visit. It is there so that what someone has to say gets kept, and reaches their family, at the hours when nobody is free to listen.

Questions

Does someone need to know anything to use it?

No. One button. Everything else is optional, and the app explains itself aloud the first time and any time it is asked to again.

What if they forget how it works?

There is a Show me how on the first screen that walks through it again, out loud, lighting up one thing at a time. It can be run any number of times.

What if they press the wrong thing?

Nothing on the screen loses a recording. Removing one only hides it, and it can be brought back.

Does it need an internet connection?

Not for recording. It works with no connection at all and nothing is lost; it catches up when the connection returns. The conversation and the natural voice need a connection.

Is it private?

Recordings are kept on the person's own device and shared only with the one family member it is set up for. There is no account, nothing is public, and nobody else is given access.

Can I get one?

This is a private project rather than a product for sale. If something here would help someone you look after, it can be built for them — shaped around their situation, on their own address, and kept entirely separate from anyone else's.