PanderoseAI/canvas-lms-academic-intent-mapping

Open Memory Standard (OMS) v1.0.0-beta

A Standardized Schema for Portable Personal AI Memories and Interconnected Context Nodes

1. Overview

The Open Memory Standard (OMS) is an open-source, vendor-neutral specification designed to eliminate data lock-in within generative AI models, contextual intelligence layers, and Personal Knowledge Management (PKM) platforms.

By standardizing personal memory payloads, applications can bridge the gap between unstructured multi-source context (such as academic dashboards, email schedules, and local markdown text) and vector databases without sacrificing user data ownership.

---

2. Platform Ingestion Objects

The baseline payload standard handles unstructured arrays exported by tier-one LLM providers and local scratchpads.

2.1 Explicit AI Memories (Platform Native Storage)

Captures explicit, atomic memory fragments stored as a flat array of unique string values.

{
  "$schema": "[https://panderose.com/schemas/v1/explicit-memory.schema.json](https://panderose.com/schemas/v1/explicit-memory.schema.json)",
  "provider": "openai",
  "version": "1.0.0",
  "extracted_at": "2026-05-24T12:22:00Z",
  "memories": [
    {
      "memory": "User is a senior student working on a satellite thermal management design project."
    },
    {
      "memory": "Prefers minimalist, infrastructure-grade, dual-color user interface layouts."
    }
  ]
}