---
title: "memory.wiki — Product Specification Synthesis"
url: https://memory.wiki/UwZWUmdo
updated: 2026-05-17T14:58:05.326Z
source: "auto-synthesis"
---
# memory.wiki — Product Specification Synthesis

> This document defines memory.wiki as a personal knowledge wiki that serves as a deployable memory layer for AI systems. It articulates a dual-door vision where users author content for AI consumption while the wiki maintains itself through LLM-powered organization.

## Key claims

- [EXTRACTED] "memory.wiki는 개인 지식 위키이면서 동시에 모든 AI에 deploy 가능한 메모리 레이어다" [doc-1]
- [EXTRACTED] The core promise is "You author. AI uses. The wiki maintains itself." [doc-1]
- [EXTRACTED] Three missing capabilities compared to current mdfy are "링크 그래프, 의미 기반 검색, LLM 자가 정리" (link graphs, semantic search, LLM self-organization) [doc-1]
- [EXTRACTED] memory.wiki explicitly rejects being "또 하나의 노트 앱" (another note-taking app) like Notion or Obsidian [doc-1]
- [EXTRACTED] The product operates through two doors: "Memory door" promising "AI가 나를 기억한다" (AI remembers me) and "Wiki door" promising "내 머릿속이 정리된다" (my mind gets organized) [doc-1]
- [EXTRACTED] Brand pillars include "One URL deploys everywhere" for cross-AI/cross-tool/cross-session usage and "Open spec, not platform lock-in" [doc-1]
- [AMBIGUOUS] The distinction between user authoring versus AI extraction is positioned as fundamental, though the specific mechanisms for maintaining this boundary are not detailed [doc-1]

## Cross-references

- The concept of "deploy everywhere" appears in both the vision statement and brand pillars, emphasizing cross-platform AI compatibility as a core differentiator [doc-1]
- Self-maintenance is referenced both in the core promise and the missing capabilities gap, suggesting LLM-powered organization is central to the product's viability [doc-1]

## Open questions / gaps

- How will the "1-click capture → URL → paste anywhere" workflow technically function across different AI platforms?
- What specific LLM-powered maintenance tasks will be automated versus requiring user oversight?
- How will the open spec/MIT license model sustain development while avoiding platform lock-in?
- What are the technical requirements for implementing semantic search and link graphs?
- How will the product handle privacy and data ownership given its "deploy everywhere" promise?

## Provenance

- [doc-1]: Defines the core product vision, positioning, and implementation gaps for memory.wiki as a self-maintaining personal wiki for AI deployment.