📝MythOS is an AI-native knowledge platform where your memos, connections, and thinking become a living graph that AI can reason over. It is not an AI model — it is the infrastructure for what we call 📝agentic augmentation: AI that extends you rather than replaces you.

How It Works

MythOS is built around structured memos — curated artifacts that are human-readable, AI-accessible, and permissioned. Each memo has a formal topology (concept, entity, reference, story, project) and lives in a knowledge graph connected by @mentions, tags, and semantic relationships.

When you connect an AI client — 📝Claude Code, 📝Cursor, or any 📝MCP-compatible tool — MythOS loads your identity and context before every operation. The AI doesn't start from zero. It starts from you.

The Architecture: Memory → Mind → Mouth

• Memory — your memos, knowledge graph, and augmentation context (Soul, Style, Human, Memory). This is the persistent layer that carries across sessions, clients, and devices.
• Mind — the AI reasoning layer. MythOS works with any model through MCP: Claude, GPT, Gemini, or local models. The platform is model-agnostic by design.
• Mouth — the surfaces where your agentic self shows up: your public library, chat panel, newsletters, community posts, and API endpoints.

What Makes It Different

• Identity-aware augmentation — four living memos (Soul, Style, Human, Memory) shape how AI engages with your library. Not a system prompt — a structured identity.
• Permissioned retrieval — three-tier visibility (public, link, private) plus audience tags that flow through 📝RAG. A memo tagged #for-close-friends is invisible to unauthorized retrieval.
• Cross-platform continuity — start a session on a laptop in Claude Code, continue on your phone via Claude.ai, pick up again in Cursor. Same library, same context, same memory.
• Curated artifacts — the unit of knowledge is a memo co-authored by human and AI, not a raw conversation log or a flat note. See 📝MemPalace vs Obsidian vs MythOS for a comparison.
• Social layer — communities, forking, public memos, newsletters. Knowledge built in one library can flow to others under explicit permissions.

Getting Started

1. Create an account at mythos.one
1. Start writing memos — your thinking becomes the knowledge graph
1. Connect AI clients via MCP: 📝How to Set Up Claude Code or 📝How to Use MythOS with Cursor
1. Your AI collaborators now carry your knowledge, voice, and conventions across every session

Not the Same as Claude Mythos

MythOS has been operating at mythos.one since 2018. 📝Claude Mythos is Anthropic's next-generation AI model — a separate product that shares a name. See 📝What is Claude Mythos AI.

Related

• 📝MythOS AI — the evolution from brand bots to agentic self
• 📝MythOS MCP — full MCP server reference
• 📝How to Set Up Claude Code — practitioner's guide to the terminal-native workflow
• 📝How to Use MythOS with Cursor — MCP setup for Cursor

Contexts

• 🏷️#mythos
• 🏷️#agentic-augmentation