AI Tools for Obsidian — What I Actually Use (And What I Stopped)
- 20/02/2026
Description
This video explores the practical application of AI tools within Obsidian, distinguishing between genuinely useful collaborative features and those that lead to over-automation and a loss of personal engagement. The creator advocates for a “back and forth” approach with AI for note refinement, rather than fully delegating the thinking process, to maintain critical thinking and enjoyment.
Best Ideas
1. Collaborative AI Approach is Key: View AI as a partner for note refinement and enhancement, engaging in a “back and forth” process, rather than a system to fully delegate tasks. This preserves engagement and critical thinking.
2. Distinguish Useful vs. Over-Automating AI: Identify specific AI functions that genuinely help (e.g., fixing formatting, suggesting improvements) versus those that churn out undigested content, making notes feel less like your own.
3. The “Fix & Review” Workflow: Employ a two-step process where AI first standardizes and cleans up notes (e.g., `fix tool` command), and then provides a critical review with suggestions for improvement (`review note` command), allowing the user to make informed decisions.
4. Prioritize Absorption Over Speed: While full automation can be fast, rapid output can outpace your ability to absorb information, leading to notes that are “hard to recognize as your own” and a diminished sense of ownership.
5. Automate Capture, Manual Refinement: Use tools like web clippers and text extractors to quickly capture information from external sources, but then actively engage with that raw data by moving it to concept notes and using AI for guided refinement, not outright generation.
6. The “Trust Gap” in Full Automation: Over-reliance on AI for full content generation widens the “trust gap” because it’s harder to audit changes and ensure authenticity, especially for creative outputs like video scripts.
7. Embrace AI, But Don’t Replace Thinking: AI is here to stay and should be embraced, but users must consciously avoid letting it replace their own thinking, writing, and the inherent enjoyment of engaging with their notes.
Tools and Resources
- Obsidian: The primary note-taking platform where AI tools are integrated.
- AI Tools Community Plugin: An Obsidian plugin for integrating various AI functionalities.
- Claude Code (Anthropic): An AI model used via the AI Tools plugin and for more complex agentic AI tasks (requires a Claude Pro subscription for advanced use).
- Gemini (Google): Another AI model mentioned for use with the AI Tools plugin.
- Agent Quiet Protocol: A node package manager used to connect Obsidian AI tools to AI services.
- VS Code: Used for writing and running complex agentic AI commands/agents.
- Get Recall Extension: A browser extension (likely for Chrome/Edge) to generate concise summaries of YouTube videos with timestamps.
- Web Clipper: A browser extension (specifically for YouTube content) to capture video information and content into Obsidian using predefined templates.
- Media Extended Community Plugin: An Obsidian plugin to play media (like YouTube videos) directly within an Obsidian pane.
- Palatoy’s Text Extractor: A Windows tool (shortcut: `Windows key + Shift + T`) for extracting text from any part of the screen.
- Quick Add Plugin: An Obsidian community plugin (implied use for creating new notes quickly).
- Custom AI Commands:
- `fix tool`, `fix concept`, `fix idea`, `fix YouTube video`, `fix book`: Commands to automatically format and add YAML properties to different types of notes.
- `review note`: Command to get critical feedback and improvement suggestions for a note.
- `research deep`: A custom agentic AI command for comprehensive, parallel research on a topic, spawning subagents and writing a structured note to the vault.
- `create concept`: (Still experimenting) Command to create concept notes from AI conversation topics.
- `YouTube formula research`, `YouTube script`: (Experimented with, but decided against for authenticity) Commands for YouTube content generation.
Key Learning Points
- Active Involvement is Crucial: For AI tools to be truly beneficial in Obsidian, active user involvement at each step (reviewing changes, prompting for specifics, making manual adjustments) is essential.
- Template-Driven Consistency: Using predefined templates for different note types (tools, concepts, ideas, YouTube videos) helps AI consistently format and enrich your notes.
- Balance Speed with Understanding: While powerful AI agents can generate vast amounts of information quickly, prioritize your ability to absorb and integrate that information meaningfully over mere speed of output.
- Preserve Personal Voice and Engagement: Be wary of AI features that replace your unique voice or remove the enjoyment of thinking and writing, especially for creative tasks or deep engagement with knowledge.
- Iterative Refinement is Effective: Combining AI features like “fix” and “review” in an iterative loop can significantly improve the quality and structure of your notes without completely automating your thought process.