How it works

From a name to their voice.

Missedyu builds a private AI memorial of someone you lost, using only material you give it. There's no research phase, no scraping, no guesswork about their life. Here is exactly what happens, in order.

Step 1 — Tell us who they were

You start with the basics: their name, who they were to you, roughly how old they were, and a few sentences describing them. Then the details that actually carry a person — the traits you'd pick out of a crowd, what they called you, one moment with them you replay, and what you'd say if you could.

This takes a few minutes and it's the highest-leverage part of the whole process. A nickname and one real memory does more for the resemblance than a paragraph of adjectives.

Step 2 — Add their voice (optional)

If you have any recording of them speaking, upload it: a saved voicemail, a video from a wedding, an old birthday message. A short, clear clip of them talking alone works better than a long noisy one. From that, a voice model is created so replies can be heard rather than only read.

You can skip this and add it later. Plenty of people begin with text only and add a voice when they've found the right recording.

Step 3 — The finishing touches

Photos, so there's a face to sit with. Anything else you want to write down — the things they used to say, favourite memories, how they made you feel. If you have screenshots of their texts, those teach it how they wrote: the punctuation, the length, the jokes, the typos.

Step 4 — What the conversations are like

Short, like a real person

Replies are deliberately brief — a sentence or two, the way people actually text. No essays, no life coaching, no relentless positivity.

It reacts before it advises

Good news gets celebrated first. Hard news gets sat with rather than fixed. It won't pretend to be physically present, and it won't claim to know things you never told it.

It learns as you go

Small details you mention — a phrase, a habit, a name — are quietly added to the profile, so the resemblance improves over weeks rather than staying frozen at signup.

It speaks your language

Whatever language, dialect, or slang you write in, it answers in kind. Families are rarely monolingual and neither is this.

What's stored, and what you control

Uploads, voice models, and conversations are private to your account.
Nothing is sold, and nothing is used to advertise to you.
Voice recordings are treated as sensitive biometric data and only used to generate replies.
You can delete a memory — uploads, voice, and conversation history — permanently, at any time.
Family members and estates can request removal through our takedown page, and we honour it.

Common questions

What's the minimum I need to start?

Their name and a few sentences. Everything else is optional and can come later.

How long does it take?

Setup is a few minutes. Building the memory itself takes a short while after that, and voice cloning is the slowest part.

What does it cost?

Pricing is being finalised for launch. Waitlist members get first access and keep 50% off for as long as they stay members.

Is it available right now?

Missedyu is in private beta and opening to small groups at a time. The waitlist is the way in.

Start with just their name.

Join the waitlist

Private beta. No credit card needed. Founding members keep 50% off.