Skip to main content

​
Overview

search_notes searches for meeting notes based on various criteria and returns metadata only (not full content). This is the first stage of the Progressive Disclosure pattern. Primary Use Cases:
  • Find meetings by content, date, or participants
  • Select specific meetings from search results
  • Verify if a note exists
Key Features:
  • Returns metadata only (title, date, participants)
  • Fast response time (0.5-1 second)
  • Token efficient (~50 tokens per note)
Does NOT Return:
  • Transcript text
  • AI summaries
  • Action items
  • Detailed documents

​
Parameters

At least one parameter is requiredYou must provide at least one of: content, createdAt, or personName
ParameterTypeRequiredDescription
contentstringOptional*Search term for note content (title or transcript)
createdAtstringOptional*Filter by creation date (ISO 8601 datetime)
personNamestringOptional*Filter by participant name or email

​
content (optional)

Search for text in note titles or transcripts. Example: 
{
  "content": "marketing"
}
Search Scope:
  • Note titles
  • Transcript content
Features:
  • Case-insensitive
  • Partial matching (e.g., “market” matches “marketing meeting”)

​
createdAt (optional)

Filter notes created on or after the specified date. Format: ISO 8601 datetime string 
YYYY-MM-DDTHH:mm:ss.sssZ
Example: 
{
  "createdAt": "2025-11-01T00:00:00Z"
}
Returns notes created on or after November 1, 2025.
Date Format RequirementsAlways include time and timezone in ISO 8601 format. Date-only formats like 2025-11-01 may not be accepted.

​
personName (optional)

Filter notes by participant name or email. Example: 
{
  "personName": "Sarah Kim"
}
Or by email: 
{
  "personName": "[email protected]"
}
Features:
  • Partial matching (e.g., “Sarah” matches “Sarah Kim”)
  • Case-insensitive

​
Response Format

​
Success Response

{
  "notes": [
    {
      "noteId": 123,
      "title": "Product Roadmap Meeting",
      "personNames": ["Sarah Kim", "John Park"],
      "createdAt": "2025-11-10T10:00:00Z"
    },
    {
      "noteId": 124,
      "title": "Weekly Team Sync",
      "personNames": ["Sarah Kim", "Emily Lee", "Mike Chen"],
      "createdAt": "2025-11-08T14:00:00Z"
    }
  ]
}
Field Descriptions:
FieldTypeDescription
noteIdnumberUnique note ID (use with other tools)
titlestringNote title
personNamesstring[]List of participant names
createdAtstringCreation date and time (ISO 8601)

​
Empty Results

{
  "notes": []
}
No matching notes found. Try relaxing search criteria.

​
Usage Examples

​
Example 1: Search by Content

Request: 
{
  "content": "product roadmap"
}
Response: 
{
  "notes": [
    {
      "noteId": 123,
      "title": "Q4 Product Roadmap Discussion",
      "personNames": ["Sarah Kim", "John Park"],
      "createdAt": "2025-11-10T10:00:00Z"
    }
  ]
}

​
Example 2: Search by Date Range

Request: 
{
  "createdAt": "2025-11-01T00:00:00Z"
}
Response: Returns all notes created on or after November 1, 2025.

​
Example 3: Search by Participant

Request: 
{
  "personName": "Sarah"
}
Response: Returns all notes where “Sarah” is a participant. Request: 
{
  "content": "marketing",
  "createdAt": "2025-11-01T00:00:00Z",
  "personName": "Sarah"
}
Response: Returns notes matching ALL criteria:
  • Contains “marketing”
  • Created after November 1, 2025
  • Has “Sarah” as participant

​
Best Practices

Begin with general searches, then add filters:
  1. Search by content only
  2. Add date filter if too many results
  3. Add participant filter for further refinement
After getting search results:
  1. search_notes - Get metadata (fast, ~50 tokens)
  2. User selects relevant notes
  3. get_note_summary - Get summary if needed (~200-800 tokens)
  4. get_note_transcript - Full details only if necessary (~3,000-5,000 tokens)
If no results:
  • Relax search criteria
  • Remove date filters
  • Use broader search terms
  • Verify notes exist in user’s account

​
Common Errors

​
Missing Parameters

Solution: Provide at least one parameter.

​
Invalid Date Format

Solution: Use ISO 8601 datetime format with timezone (e.g., 2025-11-01T00:00:00Z)

​
Token Usage

Per Note: ~50 tokens Example Scenarios:
ScenarioNotes FoundTotal Tokens
1 note1~50
5 notes5~250
10 notes10~500
50 notes (max)50~2,500
Token EfficiencyCompared to loading full transcripts (3,000-5,000 tokens per note), search_notes uses 98% fewer tokens. Always search first before loading details.

​
Next Steps

After finding notes with search_notes:

get_note_summary

Get AI-generated summary (~200-800 tokens)

get_note_transcript

Get full transcript (~3,000-5,000 tokens)

get_note_documents

Get structured data (action items, decisions)

Tools Overview

Learn about Progressive Disclosure pattern