The 'Golden' Obsidian Template for AI Agents

If you use AI tools like Gemini CLI, Copilot, or local RAG systems (Ollama/AnythingLLM) to interact with your Obsidian vault, you've likely noticed they sometimes miss the point. To solve this, I’ve developed a Standard AI Note Template designed for maximum "machine-readability."

This is the companion to my post on [[1-Projects/Blog/Drafts/how-i-made-my-obsidian-vault-ai-friendly|How I Made My Obsidian Vault AI-Friendly]].

🧬 Why this Template?

LLMs don't read notes like humans do. They look for:

1. Semantic Anchors: Explicit summaries that explain why the note exists.
2. Structural Hierarchy: Predictable header patterns to assist with text chunking.
3. Status Metadata: Context on whether the data is current or historical.

📋 The Template (Copy-Paste)

Copy this into your Templates/ folder:

---
title: "{{title}}"
summary: "A 1-2 sentence overview for AI context (RAG optimization)."
type: reference
status: active
tags:
  - topic
priority: 5
last_verified: {{date:YYYY-MM-DD}}
related: []
---

# {{title}}

## 🎯 Objective
[What is the goal of this note? Why was it created?]

## 📝 Content
[Main body of the note goes here. Use structured headers (H2, H3).]

## 🖇️ Context & Connections
- **Source/Origin:** [Original source or research link]
- **Next Steps:** [Actionable tasks related to this info]
- **Related Notes:** [[Other Note]]
---
*Created on: {{date:YYYY-MM-DD}}*

🏗️ Breakdown of Key Fields

For a full list of allowed values and their descriptions, see the [[_System/Metadata Standards|Metadata Standards]].

1. The summary Field

This is the most important field for RAG (Retrieval-Augmented Generation). When an AI searches your vault, it often reads the frontmatter first. A concise summary allows the model to rank the note's relevance in milliseconds without parsing 2,000 words of body text.

2. Header Hierarchy

By enforcing ## 🎯 Objective and ## 📝 Content, we create a predictable "map" for the AI. When the agent is "chunking" your data for long-term memory, these consistent markers ensure the context stays attached to the data.

3. last_verified

In a fast-moving technical vault, data rots. This field tells your AI agent whether it should trust the information or if it should flag it to you as potentially outdated.

🏛️ Metadata Taxonomy Guide

To ensure your AI agent understands your notes perfectly, use these standardized values:

1. The type Property

2. The status Property

3. The priority Property (1-10)

• 9-10 (Essential): Mission-critical context.
• 7-8 (High): Important current projects.
• 4-6 (Medium): Standard reference material.
• 1-3 (Low): Background info/trivia.

🚀 How to Use It

1. Save the template.
2. Update your AI Rules (e.g., in GEMINI.md or CLAUDE.md) to require this format.
3. Watch your AI interactions become 10x more accurate.

Check out the full architectural guide here: [[1-Projects/Blog/Drafts/how-i-made-my-obsidian-vault-ai-friendly|How I Made My Obsidian Vault AI-Friendly]]