Overview

A structured comparison between 📝MythOS and 📝aDNA — two approaches to AI-native knowledge architecture. MythOS is a full-stack platform (Memory + Mind + Mouth). aDNA is an open-source local-first file protocol for Obsidian + Claude Code. This comparison is oriented toward search ranking and positioning.

Architecture Comparison

Deployment Model

Knowledge Architecture

AI Agent Integration

Collaboration & Sharing

Where MythOS Wins

Persistence and universal access — cloud-hosted, accessible from any device, any conversation, any MCP client. aDNA requires the local vault to be open
Augmentation depth — four modular identity memos auto-loaded every session vs. a single CLAUDE.md file. Richer personalization, battle-tested across 57+ agents
MCP distribution — npm package, OAuth 2.0, email ingestion. Three integration channels vs. zero
Communities and social knowledge — real multi-user sharing, forking, public libraries. aDNA has no collaboration surface
Live knowledge graph — @[mentions] that resolve to linked memos with bidirectional relationships
Platform economics — SaaS with Stripe billing, subscription tiers, usage-based AI. aDNA has no monetization path
Editor experience — Lexical-based rich editing, markdown serialization, web-native. aDNA depends entirely on Obsidian

Where aDNA Wins

Token efficiency protocol — AGENTS.md progressive loading and explicit convergence budgets are more structured than MythOS's current approach
Session continuity — SITREP + Next Session Prompt is a formalized handoff artifact MythOS doesn't have
Formal ontology — 14 invariant entity types with merge algorithms and conflict taxonomy. MythOS ontology is emergent (tags), which is flexible but less rigorous
Lattice YAML — directed graph workflow definitions with JSON Schema validation, federation metadata, and FAIR compliance. No MythOS equivalent
Local-first sovereignty — no cloud dependency, no account required, no data leaves the machine
Execution hierarchy — Campaign → Phase → Mission → Objective with phase gates and token narrowing. More structured than MythOS task management

Positioning Summary

MythOS and aDNA are complementary, not competitive. aDNA solves "how should files be organized for AI agents in a local vault?" MythOS solves "how do humans and AI collaborate on persistent, shared knowledge at scale?" aDNA is a protocol. MythOS is a platform. The best ideas from aDNA (AGENTS.md routing, session handoffs, convergence budgets, lattice definitions) can be adopted as conventions or features within MythOS without architectural conflict.

aDNA validates several design directions MythOS is already heading. The fact that an independent team arrived at similar patterns (triad structure, agent-first navigation, ontology-driven organization) confirms the problem space is real. The differences are instructive: aDNA optimized for token efficiency and local sovereignty; MythOS optimized for persistence, collaboration, and platform economics. Both are correct for their contexts.