Source-backed capture
Notes, files, links, screenshots, voice, and shared content are captured as source material before being organized.
Back to Pysagor
Apple-native memory product
MemoryOS
A source-backed personal memory wiki for Apple users.
MemoryOS is being built for people who save notes, files, links, images, and decisions, then need a way to ask their own material questions later.
Problem
Personal knowledge gets scattered faster than people can reuse it.
Notes, PDFs, links, screenshots, and decisions often live in separate places. Chat-style memory can sound confident, but without sources it is hard to trust or correct.
Product answer
Turn saved material into a readable wiki that can answer with sources.
MemoryOS treats the Markdown wiki as the durable memory surface. AI can help draft answers and wiki patches, but useful changes move through a review queue before they become memory.
What is being built
The product scope is clear enough to inspect.
Markdown memory vault
The active memory layer is plain Markdown, so the product does not trap a user's memory inside one chat transcript.
Ask with citations
Ask Memory searches wiki and local context first, then presents answers with traceable source references.
Review-gated AI
AI output becomes a proposal that the user can inspect, accept, reject, or refine before memory changes.
Local-first controls
Export and delete flows are part of the product shape, keeping user agency visible in the memory system.
iCloud path, carefully stated
The app is designed for private Apple sync paths when configured, with production proof treated as a launch gate.
Current evidence
The repo points to a serious memory architecture, not a generic notes app.
The product direction is local-first and review-gated: MemoryOS keeps raw sources, a local database, and a user-readable Markdown vault as the core memory layer. iCloud sync is part of the implementation path, but production CloudKit and two-device proof remain a separate release gate.
The vault shape is explicit.
MemoryOS uses raw source material, wiki Markdown, schema data, SQL records, search copies, review state, and audit metadata.
AI has a governance boundary.
Source repo evidence keeps AI suggestions reviewable and avoids direct, silent rewrites to active memory.
Release proof is still separate.
RevenueCat and App Store preparation exist in the source repo, but public release still depends on TestFlight, purchase, restore, and Apple approval proof.
Related resources
Legal pages are already routed on the landing site.
The MemoryOS privacy and terms pages describe local storage, iCloud and CloudKit dependency, RevenueCat subscriptions, Apple services, and acceptable use.
Product questions
Direct answers for search and review.
What is MemoryOS?+
MemoryOS is an Apple-native personal memory app in development for source-backed recall, reviewable AI suggestions, and a local Markdown wiki.
Is MemoryOS available now?+
No. MemoryOS is in development and has not been published yet, so this page describes the product direction and release-stage context.
How does MemoryOS handle AI suggestions?+
MemoryOS is designed around review-gated AI suggestions, so proposed memory or wiki changes can be inspected before they become part of the user's memory layer.
Next step
Want to understand MemoryOS before launch?
Contact Pysagor for the product direction, privacy posture, and current release-stage context.