Skip to main content

​
Overview

get_note_documents returns structured information from meetings. Primary Use Cases:
  • Extracting action items
  • Reviewing section summaries
  • Getting structured meeting notes
  • Planning follow-up tasks
Key Features:
  • Structured data (JSON format)
  • Action item focused
  • Organized by sections
Included Information:
  • Action items (assignee, due date)
  • Section summaries
  • Key decisions
  • Other structured data

​
Parameters

ParameterTypeRequiredDescription
noteIdnumberYesNote ID (positive integer)

​
noteId (required)

The unique ID of the note to retrieve. Example: 
{
  "noteId": 123
}
How to Get Note ID:
  1. Use search_notes to find meetings
  2. Use the noteId from search results

​
Response Format

​
Success Response

{
  "documents": {
    "summary": "Q4 Marketing Strategy Meeting",
    "actionItems": [
      {
        "task": "Draft marketing plan",
        "assignee": "John Park",
        "dueDate": "2025-11-20"
      },
      {
        "task": "Review budget proposal",
        "assignee": "Sarah Kim",
        "dueDate": "2025-11-15"
      }
    ],
    "sections": [
      {
        "title": "Current Performance",
        "content": "Website traffic increased 30%, conversion rate needs improvement"
      },
      {
        "title": "Q4 Goals",
        "content": "20% revenue increase, focus on new customer acquisition"
      }
    ],
    "decisions": [
      "Approved $50,000 marketing budget",
      "Decided to adopt influencer marketing"
    ]
  }
}
Field Descriptions:
FieldTypeDescription
documentsobjectStructured document container
documents.summarystringOne-line meeting summary
documents.actionItemsarrayList of action items
documents.actionItems[].taskstringTask description
documents.actionItems[].assigneestringPerson responsible
documents.actionItems[].dueDatestringDue date (ISO 8601)
documents.sectionsarraySection contents
documents.sections[].titlestringSection title
documents.sections[].contentstringSection content
documents.decisionsarrayKey decisions made

​
Usage Examples

​
Example 1: Extract Action Items

AI Request: 
Show me the action items from note #123
MCP Call: 
{
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "name": "get_note_documents",
    "arguments": {
      "noteId": 123
    }
  },
  "id": 1
}
AI Response: 
Q4 Marketing Strategy Meeting Action Items:

1. John Park
   - Task: Draft marketing plan
   - Due: 2025-11-20

2. Sarah Kim
   - Task: Review budget proposal
   - Due: 2025-11-15

3. Emily Lee
   - Task: Competitor analysis report
   - Due: 2025-11-18

​
Example 2: Review Key Decisions

AI Request: 
What decisions were made in this meeting?
Workflow: 
1. search_notes(recent meeting)
   → noteId 123

2. get_note_documents(noteId: 123)
   → Extract decisions field
AI Response: 
Key Decisions:

- Approved $50,000 marketing budget
- Decided to adopt influencer marketing
- Focus on social media advertising
- Set Q4 revenue target at 20% increase

​
Example 3: Section-by-Section Breakdown

AI Request: 
Organize the meeting content by sections
AI Response: 
Q4 Marketing Strategy Meeting (By Section):

Current Performance
- Website traffic increased 30%
- Conversion rate needs improvement
- Reached 20K social media followers

Q4 Goals
- 20% revenue increase
- Acquire 1,000 new customers
- Improve brand awareness by 40%

Strategy
- Adopt influencer marketing
- Strengthen content marketing
- SEO optimization

Budget
- Total $50,000 approved
- Advertising: $30,000
- Content: $15,000
- Other: $5,000

​
When to Use

Follow-up Tasks Needed
  • Tracking action items
  • Checking assignee responsibilities
  • Managing deadlines
Structured Information
  • Understanding section content
  • Organizing key decisions
  • Systematic meeting notes
Project Management
  • Extracting tasks
  • Schedule management
  • Team member assignments
Decision Tracking
  • Confirming decisions made
  • Approved budgets/plans
  • Policy changes
Only Need Simple Summary → Use get_note_summary insteadNeed Full Conversation → Use get_note_transcript insteadExploration Phase → Use search_notes instead

​
Performance

​
Response Time

Meeting LengthAverage ResponseMaximum Response
30 minutes1 second3 seconds
1 hour1.5 seconds5 seconds
2 hours2 seconds7 seconds

​
Token Usage

Meeting LengthDocuments TokensTranscript TokensSavings
30 minutes~300-600~1,500-2,50070%
1 hour~500-1,000~3,000-5,00075%
2 hours~800-1,500~6,000-10,00080%
Structured Data AdvantageDocuments provide structured JSON format, making it easy to extract specific information like action items or decisions without parsing unstructured text.

​
Real-World Scenarios

​
Scenario 1: Weekly Task Compilation

User Request: 
Compile all action items from this week's meetings
AI Workflow: 
1. search_notes(createdAt: "this week")
   → Found 5 meetings

2. Call get_note_documents for each meeting
   → Extract actionItems

3. Group by assignee

Result:
Sarah Kim:
- Review budget proposal (11/15) - Q4 Marketing Meeting
- Design review (11/18) - Product Meeting
- Team performance report (11/20) - Weekly Sync

John Park:
- Marketing plan draft (11/20) - Q4 Marketing Meeting
- Server performance improvement (11/22) - Tech Meeting
...

​
Scenario 2: Budget Approval Tracking

User Request: 
Find all approved budgets from last month
AI Workflow: 
1. search_notes(createdAt: "last month", content: "budget")
2. Call get_note_documents for each note
3. Search for "approved" in decisions field

Result:
Budget Approvals:
- Q4 Marketing: $50,000 (Approved 11/10)
- New Hires: $120,000 (Approved 11/05)
- Server Upgrade: $30,000 (Approved 10/28)
Total: $200,000

​
Limitations

​
Current Limitations

AI-Generated DataDocuments are AI-generated with ~85-90% accuracy. For critical information, verify against the full transcript using get_note_transcript.
Known Limitations:
  1. Fixed Document Structure
    • Only returns predefined fields
    • Custom structures not supported
  2. Action Item Recognition
    • Only recognizes explicit task assignments
    • Implicit tasks may be missed
  3. Accuracy Considerations
    • ~85-90% accuracy for AI-generated content
    • Important details should be verified with transcript

​
Future Improvements

FeatureTargetDescription
Custom Templates2025 Q2Generate documents with custom structures
Accuracy Improvement2025 Q2Model improvements targeting 95% accuracy
Additional Fields2025 Q3Risks, issues, notes, etc.

​
Common Errors

​
Missing Note ID

Solution: Provide the noteId parameter.

​
Invalid Note ID

Solution: Use a positive integer (e.g., 123, not "123" or -1).

​
Note Not Found

Possible Causes:
  • Note doesn’t exist
  • Note has been deleted
  • You don’t have access to this note
Solution: Verify the note ID using search_notes.

​
No Documents Available

Processing RequiredIf a note has just been created, documents may not be available yet. Wait for processing to complete (usually 1-2 minutes after meeting ends).

​
Next Steps

After reviewing documents:

get_note_transcript

Get full conversation if more detail needed

get_note_summary

Get quick summary for overview

search_notes

Find other related meetings

Tools Overview

Learn about all available tools