Load Meeting

Fetch meeting details, attendees, transcript, and insights

The Load Meeting node fetches comprehensive meeting data including details, attendees, call recording, transcript, and AI-generated insights. It transforms a meeting ID into rich, actionable data for your workflow.

Overview

Property	Value
Category	Loader
Node ID	`ds.loadMeeting.perItem.in1.success1.error1`
Input Ports	1
Success Outputs	1
Error Outputs	1
Execution Mode	Per-item

Configuration

Parameter	Type	Required	Description
`meeting_id`	CEL Expression	Yes	The meeting plan ID to load

Meeting ID Sources

The meeting ID typically comes from the trigger:

cel

trigger.meetingPlanId

Or from upstream node output:

cel

json.meetingId

Input Schema

This node expects to receive any data from upstream. The only requirement is that the meeting_id parameter can be resolved.

Output Schema

Success Output

json

{
  "meeting": {
    "id": "mp_abc123def456",
    "title": "Q1 Planning Review",
    "startTime": "2024-01-15T14:00:00Z",
    "endTime": "2024-01-15T15:00:00Z",
    "location": "https://zoom.us/j/123456789",
    "attendees": [
      {
        "name": "John Smith",
        "email": "john@example.com",
        "response": "accepted"
      },
      {
        "name": "Jane Doe",
        "email": "jane@example.com",
        "response": "accepted"
      }
    ]
  },
  "callRecording": {
    "id": "cr_xyz789",
    "createdAt": "2024-01-15T15:05:00Z",
    "duration": 3600,
    "transcriptSummary": "The team discussed Q1 objectives, focusing on three main areas: revenue targets, product roadmap, and team expansion. Key decisions were made about prioritizing the enterprise segment.",
    "transcript": "John: Welcome everyone to our Q1 planning review...",
    "keyStatements": [
      {
        "speaker": "John Smith",
        "text": "We need to focus on enterprise accounts this quarter",
        "timestamp": "00:05:23",
        "sentimentScore": 0.8,
        "classification": "decision"
      },
      {
        "speaker": "Jane Doe",
        "text": "The timeline for the new feature is aggressive",
        "timestamp": "00:12:45",
        "sentimentScore": -0.2,
        "classification": "concern"
      }
    ],
    "insights": {
      "topics": ["Q1 planning", "revenue", "enterprise"],
      "sentiment": "positive",
      "actionItems": [
        "Follow up with enterprise prospects",
        "Schedule product roadmap review"
      ]
    }
  }
}

Output Fields

`meeting` Object

Field	Type	Description
`id`	string	Meeting plan ID
`title`	string	Meeting title from calendar
`startTime`	string (ISO 8601)	Scheduled start time
`endTime`	string (ISO 8601)	Scheduled or actual end time
`location`	string	Meeting location or join URL
`attendees`	array	List of meeting attendees

`attendees` Array Items

Field	Type	Description
`name`	string	Attendee's display name
`email`	string	Attendee's email address
`response`	string	Calendar response: "accepted", "declined", "tentative", "none"

`callRecording` Object

Field	Type	Description
`id`	string	Recording ID
`createdAt`	string (ISO 8601)	When recording was processed
`duration`	number	Recording duration in seconds
`transcriptSummary`	string	AI-generated summary of transcript
`transcript`	string	Full transcript text
`keyStatements`	array	Notable statements extracted by AI
`insights`	object	AI-generated meeting insights

`keyStatements` Array Items

Field	Type	Description
`speaker`	string	Who said it
`text`	string	What was said
`timestamp`	string	When in the recording (HH:MM:SS)
`sentimentScore`	number	Sentiment (-1 to 1, positive = positive sentiment)
`classification`	string	Type: "decision", "question", "concern", "action", etc.

Error Output

json

{
  "error": "MEETING_NOT_FOUND",
  "message": "Meeting with ID mp_abc123 not found",
  "meetingId": "mp_abc123"
}

Error Code	Description
`MEETING_NOT_FOUND`	No meeting exists with this ID
`NO_RECORDING`	Meeting exists but has no recording yet
`PERMISSION_DENIED`	User doesn't have access to this meeting
`TRANSCRIPT_PENDING`	Recording exists but transcript is still processing

Examples

Basic Example: Load After Trigger

Load meeting data immediately after trigger fires:

text

[Event Trigger: MEETING_ENDED]
             ↓
       [Load Meeting]
   meeting_id: trigger.meetingPlanId

Configuration:

meeting_id: trigger.meetingPlanId

Using Meeting Data

After loading, access meeting data in downstream nodes:

AI Prompt User Message:

liquid

Summarize this meeting:

Title: {{ json.meeting.title }}
Attendees: {% for a in json.meeting.attendees %}{{ a.name }}{% unless forloop.last %}, {% endunless %}{% endfor %}

Transcript:
{{ json.callRecording.transcript }}

Slack Message:

liquid

📋 *Meeting Recap: {{ json.meeting.title }}*

*Attendees:* {% for a in json.meeting.attendees %}{{ a.name }}{% unless forloop.last %}, {% endunless %}{% endfor %}

*Summary:*
{{ json.callRecording.transcriptSummary }}

*Key Decisions:*
{% for stmt in json.callRecording.keyStatements %}{% if stmt.classification == "decision" %}
• {{ stmt.text }} ({{ stmt.speaker }})
{% endif %}{% endfor %}

Advanced Example: Handle Missing Transcript

Some workflows need the transcript, which may not be immediately available:

text

[Event Trigger: MEETING_ENDED]
             ↓
[Wait: 5 minutes]  ← Give time for transcript processing
             ↓
       [Load Meeting]
       ↓           ↓
    Success      Error → [Slack: "Transcript not ready"] → [Sink]
       ↓
  [AI Analysis]
       ↓
   [Slack Post]

Example: Extract Specific Attendee

Get the first external attendee (non-company email):

AI Prompt:

liquid

{% assign external = nil %}
{% for a in json.meeting.attendees %}
  {% unless a.email contains "@ourcompany.com" %}
    {% assign external = a %}
    {% break %}
  {% endunless %}
{% endfor %}

{% if external %}
Contact name: {{ external.name }}
Contact email: {{ external.email }}
{% else %}
No external attendees found.
{% endif %}

Best Practices

1. Always Use Trigger's Meeting ID

For meeting-triggered workflows, always use the trigger's ID:

✅ Correct:

cel

trigger.meetingPlanId

❌ Incorrect:

cel

"mp_hardcoded123"  // Never hardcode meeting IDs

2. Handle Error Output

Connect the error output for graceful handling:

text

[Load Meeting]
├─ Success ─→ [Continue Processing]
└─ Error ───→ [Log Error] → [Sink]

3. Wait for Transcript When Needed

If your workflow depends on transcript content, add a brief wait:

text

[Trigger] → [Wait: 2-5 min] → [Load Meeting]

Transcript processing typically completes within minutes of meeting end.

4. Check for Missing Data

Not all fields are always populated:

liquid

{% if json.callRecording %}
  Summary: {{ json.callRecording.transcriptSummary }}
{% else %}
  No recording available for this meeting.
{% endif %}

Common Issues

"Meeting not found" error

Symptoms: Error output fires with MEETING_NOT_FOUND

Solutions:

Verify the meeting ID is correct (trigger.meetingPlanId)
Check the meeting exists in Decision Site
Ensure the workflow's organization matches the meeting's organization

Empty transcript

Symptoms: callRecording.transcript is empty or null

Solutions:

Add a Wait node before Load Meeting (2-5 minutes)
Check if the meeting was recorded
Verify transcription is enabled for your account

Missing attendees

Symptoms: attendees array is empty

Solutions:

Check calendar event has attendees
Verify calendar sync is working
Attendees must be on the calendar invite

Stale data

Symptoms: Data doesn't reflect recent changes

Solutions:

Load Meeting fetches fresh data each execution
If data seems stale, check the source (calendar, recording system)

Event Trigger - Provides meeting ID
AI Prompt - Process transcript with AI
AI Agent - Complex analysis of meeting data

Technical Details

Data Sources

Load Meeting aggregates data from multiple sources:

Meeting details: Calendar integration
Attendees: Calendar event participants
Recording: Decision Site recording system
Transcript: Speech-to-text processing
Insights: AI analysis pipeline

Processing Time

Typical availability after meeting ends:

Meeting details: Immediate
Attendees: Immediate
Recording: 1-2 minutes
Transcript: 2-5 minutes
Insights: 3-5 minutes

Rate Limits

Load Meeting operations are subject to internal rate limits:

Typical limit: Generous for normal usage
If you see rate limit errors, add delays between loads