See It In Action
Watch how AI assistants intelligently interact with your Obsidian vault. These examples show real conversations and outcomes.
AI Desktop Tool - MCP-Obsidian Active
Connected
You
Add the equation for energy-mass equivalence to my physics notes
AI
AI is thinking...
MCP-Obsidian: Using patch_note... Request:
{
"path": "Physics/Relativity.md",
"oldString": "## Energy and Mass",
"newString": "## Energy and Mass\n\nE = mc²"
}Response:
{
"success": true,
"path": "Physics/Relativity.md",
"message": "Successfully replaced 1 occurrence",
"matchCount": 1
}Done! Added Einstein's equation to your notes. Only the specific section was updated - no full file rewrite needed.
Technical Details
MCP server used patch_note for efficient partial update
Multi-line text replacement performed atomically
Frontmatter and formatting preserved perfectly
10x faster than rewriting entire file
replaceAll defaults to false — fails safely if multiple matches found
You
Create a quick note about today's meeting
AI
AI is thinking...
MCP-Obsidian: Using write_note... Request:
{
"path": "Meetings/Team Sync.md",
"content": "# Team Sync\n\n- Discussed Q1 goals\n- Action items assigned"
}Response: Successfully wrote note: Meetings/Team Sync.md (mode: overwrite) Done! I've created your meeting note.
Technical Details
MCP server wrote the complete note to disk
AI generated the structured content
File created atomically with proper formatting
Ready to open and edit in Obsidian
Supports overwrite, append, and prepend modes
You
Read all my book club notes and give me a summary
AI
AI is thinking...
MCP-Obsidian: Using read_multiple_notes... Request:
{
"paths": [
"Reading/The Phoenix Project.md",
"Reading/Atomic Habits.md",
"Reading/Deep Work.md"
],
"prettyPrint": true
}Response:
{
"ok": [
{
"path": "Reading/The Phoenix Project.md",
"content": "...",
"frontmatter": {...}
}
],
"err": []
}Done. I've read all 3 notes. Here's a summary: Common themes across these books: systems thinking, habit formation, and focus. All emphasize the importance of intentional work practices and continuous improvement.
Technical Details
MCP server read multiple files in one request (max 10)
Returns ok (successful) and err (failed) arrays
AI analyzes across multiple documents
More efficient than reading files one-by-one
prettyPrint defaults to false for minimal token usage
You
Update the status and add tags to my project planning note
AI
AI is thinking...
MCP-Obsidian: Using update_frontmatter... Request:
{
"path": "Projects/Website Redesign.md",
"frontmatter": {
"tags": [
"project",
"web-design",
"priority-high"
],
"status": "in-progress",
"created": "2025-01-15",
"updated": "2025-01-20"
}
}Response: Successfully updated frontmatter for: Projects/Website Redesign.md Done! I've updated the frontmatter with the new status and tags.
Technical Details
MCP server safely updated YAML frontmatter
Merges with existing frontmatter by default (merge: true)
Updates are performed atomically
Note content remains untouched
YAML is validated before writing to prevent corruption
You
Search for "React hooks" in my notes
AI
AI is thinking...
MCP-Obsidian: Using search_notes... Request:
{
"query": "React hooks",
"limit": 10,
"prettyPrint": true
}Response:
[
{
"p": "Development/React Best Practices.md",
"t": "React Best Practices",
"ex": "...State **React hooks** provide...",
"mc": 8,
"ln": 42
},
{
"p": "Learning/Modern JavaScript.md",
"t": "Modern JavaScript",
"ex": "...useEffect are common **React hooks**...",
"mc": 3,
"ln": 156
}
]Done. Found 2 notes with 11 total matches across your vault.
Technical Details
MCP server performed full-text search across vault
Token-optimized response: p=path, t=title, ex=excerpt, mc=matchCount, ln=lineNumber
Returns 21-char context excerpts around matches
AI can then read specific files for more details
prettyPrint defaults to false for minimal token usage
Ready to experience this level of AI-powered note management?
Get Started Now