Mori · My Atlas

an AI companion that shows its work

Everything Mori believes about you, in a book you can edit.

Most AI companions quietly build a picture of you and never show it. Mori keeps its picture in My Atlas — every belief labeled as your words or Mori's guess, backed by dated evidence, rated for confidence, and yours to confirm, correct, or erase.

4

confirmed patterns

6

open hypotheses

1

corrected by you

4

open questions

the two-minute proof

Catch Mori being wrong — then watch it change its mind in front of you.

1

Ask

Open Ask Mori and tap “What have you learned about me?”

2

Inspect

Open the evidence behind Mori's guess that “social interactions drain you.” Three events — all large groups.

3

Correct

Tell Mori the truth: large group events drain you, but time alone with a close friend helps.

4

Watch it update

Mori owns the mistake, shows exactly what changed in your Atlas — and asks better questions from now on.

Start the flow →

anatomy of a memory

Every memory card shows its receipts.

  • Who said it: “You told me” vs. “Mori's hypothesis” — observation and interpretation never blur.
  • How sure: a confidence level Mori actually honors — low-confidence guesses are held, not acted on.
  • The evidence: dated observations you can open, from conversations you can trace.
  • Why it's kept: every entry explains what it's for.
  • Your controls: confirm, correct, keep observing, hide, or delete — right on the card.

a real card from the sample archive ↓

You confirmed thisMori's inferencehigh confidence

A short walk after lunch reliably steadies you on stressful days.

First noticed by Mori on 29 April; you confirmed it on 6 May — “the walk thing is real, by the way.”

Source: Lunch check-in · 6 May 2026Period: April – July 2026

Why Mori keeps this: A confirmed, repeatable steadier — one of the first things Mori reaches for when you say a day is going sideways.

memory controls

Four connected spaces, one transparent memory