Skip to content
Iterus

Legacy You · pre-launch

Legacy You: A Digital Legacy Designed for Seniors

Legacy You is an Iterus product: a platform where a person records messages for their loved ones — voice, video, practical information — delivered only after their death, never earlier and never to anyone else. It is designed senior-first: voice control, large touch targets and passwordless sign-in. The native mobile app and the web are built; the product is pre-launch.

What problem does Legacy You solve?

People have no reliable way to leave their last messages to loved ones so that they arrive at the right moment — after death, not before — and only reach the people they are meant for. And the tools that attempt it tend to be unusable for seniors.

Legacy You starts from the other end: the senior first. The user records messages and names recipients and trustees; a regular "check-in" confirms they are fine, and when they stop responding, the app steps up its alerts and triggers delivery only after the trustees confirm the death. Primarily for active seniors and people with a serious diagnosis; the expectation is that the "sandwich generation" will set it up for their parents.

What did we build?

  • Native mobile app for iOS and Android (React Native, Expo) — native camera, audio and video recording, biometric unlock, keys in the phone's hardware key store, haptics. Built with EAS and verified on a real device in August 2026; it is not a website wrapped in an app (PWA).
  • Web application (Next.js) — marketing site, dashboard, portals for recipients, trustees and heirs, public pages for funeral services.
  • Supabase backend — edge functions, row-level data isolation (Row Level Security), a scheduled pipeline for escalation and delivery, and a regular health check that watches it.
  • Messages and vault — text, audio and video with end-to-end encryption, time capsules, funeral messages with a one-time code, a digital vault the server cannot read (zero-knowledge, AES-256-GCM), a family tree and heir access.
  • Check-in and trustees — confirmation by app, e-mail or voice, alerts stepped up over several levels with a pause option; trustees decide by consensus through their own scoped portal; delivery to recipients by e-mail, verified into a real inbox.
  • Voice and accessibility — voice control (Whisper → Claude → ElevenLabs), an audio guide on every screen, voice banking and a mode for people who may lose their voice, a three-level caregiver mode. Accessibility is a CI requirement, not a conformance statement: an automated check of the public part (axe) with no findings, plus touch-target, focus and contrast checks on every protected screen; manual screen-reader tests are still ahead.
  • AI Memory — search over the deceased person's content for the bereaved, only with explicit consent given during their lifetime. It works only with content the user explicitly released for the bereaved while alive; everything else stays encrypted.

The numbers

unit tests (web)
1,028
unit tests (mobile app)
418
end-to-end tests (sign-in and public pages)
218
web routes (site, dashboard, portals)
101
native mobile-app screens
85
edge functions (Supabase)
20
database migrations
55
commits since November 2025
549

The product is pre-launch; the numbers above describe the scope of what is built and tested. The delivery pipeline is verified in production from end to end.

How did we work?

We work AI-native: delivery is accelerated with Claude Code and Codex, and quality is secured by automated tests, independent code review and specialists on demand.

The human — the project's technical lead — made the product and security decisions: senior-first principles, the encryption architecture and the zero-knowledge vault, trustee consensus, what stays private even after death. He ran everything the AI holds no keys to — the database project and its access keys, hosting and domains, DNS for e-mail delivery, key rotation, native builds and the app-store accounts. Every write to production and every irreversible operation went through his explicit approval. He verified the result on a real phone, with live sign-ins on production and by checking delivery into inboxes. He also covered the legal and compliance layer — vendor agreements, biometric consents and legal texts with an external legal review.

Claude Code, orchestrating other models for review, implemented against tests (test → code → measurement), wrote automated checks and end-to-end probes, migrations and edge functions; measured instead of assuming (timeout timings, contrast ratios, mutation tests); had its own changes critically reviewed by several models before every push and verified each finding by running it against the code.

AI writes the code under human direction; the human decides, runs production and verifies the results against reality — and it is exactly those checks that have repeatedly caught defects the automation let through.

What's next?

The first version's code is finished and verified in production. Launch waits on steps outside the code: legal review, testing with seniors and the changes that follow, manual screen-reader tests and app-store review. We'll update the status here when it changes.

A native mobile app, accessibility as a requirement and encryption with keys on the user's side are the same things we can build for your product: see web applications, or simply get in touch.