Stop re-explaining yourself to your AI.
Write the way you already do — notes, journals, articles, lists. Every note becomes a Living DocumentLiving DocumentA living document — also known as a dynamic document — is a continually edited and updated document, in contrast to "dead" or "static" documents that are written once and locked against the evolving...#notecard#mythos-lexiconRead the memo → your AI reads, cites, and updates with you, from MythOS, ChatGPT, or Claude. One Single Source of TruthSingle Source of TruthIn information systems design, the curation of a single source of truth (SSOT) is the practice of structuring information such that every element is stored exactly once. Any possible linkages to an...#mythos-deck#mythos-lexiconRead the memo →, shared on your terms.
Free to start · No credit card required
Your library, one click from every AI.
You already have a memory. It's just scattered across Notion pages, voice memos, half-written drafts, and every AI chat you've ever had.
Reading list for the memory & cognition essay:
- The Extended Mind — Annie Murphy Paul #memory
- How to Take Smart Notes — Sönke Ahrens
- The Maniac — Benjamín Labatut · about von Neumann
- 01Monday: you write a note in MythOS, like you would anywhere. Books become @mentions, topics become #tags — without your hands leaving the keyboard.
- 02Wednesday, in ChatGPT or Claude: “add The Maniac to my book research.” Your AI updates the same note, over MCP.
- 03Whenever you return: one current copy. No stale duplicates, no “where did I put that.”
Your library, connected
This is what your thinking looks like after it compounds.
Every @mention and #tag becomes a connection. Over time, your library becomes a map of how you think.
# Tuesday, in Claude Code
> Get everything we know about the drafting loop.
⏺ mythos - search_memos (MCP)(tags: "drafting-loop")
⎿ 4 memos loaded as context
> We changed the review step — update the loop.
⏺ mythos - update_memo (MCP)(memoId: "drafting-loop")
⎿ Updated — every future session sees this
# Thursday, in Cursor
> Run the drafting loop on this doc.
⎿ Loop loaded — including Tuesday's change
You keep a CLAUDE.md in every project. The same context blocks, duplicated fifty times — and when one changes, you update it fifty times.
In MythOS that block is one living memo, indexed by hashtag, linked from anywhere. Any session invokes it on demand — no token bloat — and any agent can update it once, for every project at once. Infinite CLAUDE.mds, zero duplication.
Connect your tools via MCPWrite once. Remember forever.
Markdown editor with
@mentions
,
#hashtags
, collapsible headings, and images that paste in like email. Your memos link to each other automatically. Your library builds itself.
Notes on the memory essay — @The Extended Mind
Annie Murphy Paul — thinking happens beyond the brain: notebooks, spaces, other people.
#memoryBuild your library with your AI, not around it.
MythOS is designed to be co-authored. Ask Claude to research a topic — it writes the memo and files it for you. Use Cursor, Windsurf, or any MCP-native app, and your library follows. Every conversation ends with something permanent.
Connect via MCP →> Research spaced repetition for the essay — and file it.
⏺ mythos - create_memo (MCP)(title: "Spaced Repetition")
⎿ Filed to your library — permanent
Review schedules beat re-reading — the forgetting curve flattens each time you recall, not each time you read.
One library. Every workflow.
Drop a memo into any automation. Zapier reads from MythOS. Your n8n flow pulls your notes before firing. Your agent loads your library as context before replying. Your memory becomes the single source of truth the rest of your stack can actually reference.
See integrations →Talk to what you've written.
Ask your library a question and it answers in your own words — with a citation to the exact memo where you wrote it. Search by meaning, not keywords. Your past thinking, on demand.
Try the chat →Your library, out loud.
Publish any memo with one toggle. Readers get your prose, your mentions, your tags, and a chat panel that speaks in your voice — grounded in your own memos. Every memo is a live page. Every library is a live conversation.
See a public memo →Notes don't think. Networks do.
A #memory system that survives every model migration, grounded in @the-portable-context doctrine. Ideas become nodes; nodes become navigable.
One source. Three channels.
Turn any public memo into a discussion space — upvote, comment, fork. Send it as a newsletter issue to your subscribers. Get surfaced in search — Google and every AI indexer, thanks to JSON-LD and
/llms.txt
. Your audience doesn't come to the platform; the platform goes to them.
Why MythOS exists.
Start Here
This page you're reading is a memomemoThe word memo is short for a memorandum; (1) an informal record or written reminder, (2) a record of communication that contains a directive, advisory, or informative matter. The Memo Emoji prefaces...#mythos-lexiconRead the memo → — one note in a library of thousands I've been writing since 2017. Everything underlined is alive: hover it, click it, follow it as deep as you care to go.
I built MythOS because my thinking was scattered across notebooks, apps, and drafts that couldn't talk to each other — and then AI arrived, and the scattering got expensive. Every new session started from zero. Every project meant re-explaining myself. So I built the alternative: a Single Source of TruthSingle Source of TruthIn information systems design, the curation of a single source of truth (SSOT) is the practice of structuring information such that every element is stored exactly once. Any possible linkages to an...#mythos-deck#mythos-lexiconRead the memo → where every note is a Living DocumentLiving DocumentA living document — also known as a dynamic document — is a continually edited and updated document, in contrast to "dead" or "static" documents that are written once and locked against the evolving...#notecard#mythos-lexiconRead the memo → — written once, linked to everything it touches, and kept current by me and the AI agents I work with, from whatever app they live in.
That's the whole idea. You write the way you already write — journals, articles, lists, half-formed notes. Your library organizes itself as you go, like a ZettelkastenZettelkastenZettelkasten is a method of note-taking developed by German sociologist Niklas Luhmann, designed to systematically organize and interconnect personal knowledge. Unlike traditional note-taking...#what-is-mythosRead the memo → that builds its own index. And every AI you use can read it, cite it, and update it with you, so nothing you think gets lost between tools, sessions, or years.
I wrote fifty million words into this system before ChatGPT launched. It was worth it then. It's indispensable now.
Make one of your own — start with a single note. That's how this one started.
Contexts
Trusted by thinkers from
“First you upload your brain, then you start making edits.”
John ZdanowskiCFO, Second Life“Because of the way it's built, your story can't be hijacked.”
Jim GoodmanTicketmasterBrian Swichkow started writing his library in 2017. Fifty million words before ChatGPT launched.
Start remembering everything.
Free to start · No credit card required
Simple tiers. Yours forever.
Three tiers, no surprises. Export everything anytime.
- Unlimited memos + daily journal
- Media uploads (5MB/file, 500MB total)
- Voice transcription
- Custom feed and memo URLs
- Everything in Scribe
- Public memo publishing
- Unlimited collaborators
- Agent API + MCP + /llms.txt
- Email ingestion
- Everything in Scholar
- Custom domain
- Community spaces + moderation
- AI agents trained on your library
- Adaptive AI as a Service
Before you ask.
Is this related to Anthropic's Claude "MythOS"?—
No. We've been MythOS since 2017 — years before Anthropic existed. If anything, it's the other way around.
What if I want to leave?—
Export everything as markdown anytime, including media. Your library is yours; we're just hosting it.
Do I need to know markdown?—
No. The editor works like any rich-text tool — bold, lists, checkboxes, images. Markdown is what we save underneath so it stays portable forever.
What AIs does this work with?—
Claude, ChatGPT, Gemini, Grok, Perplexity — anything that can speak HTTP or MCP. We publish an llms.txt and an agent-scoped API. Your next model will work too.
Can I bring my own AI key?—
Yes. Free and Pro both support BYO keys for OpenAI or Anthropic. Pro also includes hosted AI if you'd rather not manage keys.
Is my library private by default?—
Yes. Every memo starts private. You choose what to make public, unlisted, or shared with specific contacts via permission tags.
Not ready yet? Take the ideas with you.
Occasional letters on building a memory your AI can actually use. No spam, unsubscribe anytime.