**Option A: Use Existing Worker**
- Good for general-purpose automation
- Can handle multiple webhook types**Option B: Create Dedicated Worker** (Recommended)
- Specialized for specific webhook events
- Tailored instructions for consistent handling
- Easier to track and manage

    Example: "Meeting Transcription Processor"
    Purpose: Handle meeting recording webhooks from Fathom

    1. Go to worker → Integrations → Browse Apps
    2. Search for the service (e.g., "Fathom")
    3. Click "Add" → "Create new connection"
    4. Get API key from service:
       - Log into third-party service
       - Find API/Developer settings
       - Generate new API key
       - Copy key
    5. Paste API key into Utari
    6. Click "Connect"
    7. Enable relevant tools

**Method 1: From Worker**

    1. Go to worker → Triggers tab
    2. Click "Create new trigger"
    3. Select "Webhook trigger"

**Method 2: From Triggers Dashboard**

    1. Go to Triggers → Add Trigger
    2. Select "Webhook trigger"
    3. Choose the worker

Define what your agent should do when a webhook is received**Basic Structure**:

    Process the following [event type].    Extract:
    - [Key information 1]
    - [Key information 2]
    - [Key information 3]    Then:
    1. [Action step 1]
    2. [Action step 2]
    3. [Action step 3]    Do not ask for confirmation—execute automatically.

**Available Variables**:Use these placeholders in your template - they'll be replaced with actual webhook data:- `{{payload}}` - Full webhook payload as JSON
- `{{payload.field}}` - Specific field from payload (e.g., payload.title)
- `{{payload.nested.field}}` - Nested field access
- `{{timestamp}}` - When the webhook was received
- `{{trigger_name}}` - Name of this trigger**Example Template**:

    Process the following meeting recording:    Title: {{payload.meeting_title}}
    Recording URL: {{payload.recording_url}}    Please summarize the key points and action items.

**Model Selection**:Choose which model to use for processing webhooks- **Utari Basic** (Recommended for most webhook processing tasks)

You'll receive:
- **Webhook URL**: Starts with `https://utari.up.railway.app/v1/triggers/...`
- **Secret Token**: One-time display (if enabled)**IMPORTANT: Copy both immediately!**Example webhook URL:

    https://utari.up.railway.app/v1/triggers/373a2...

The secret token should be included in the `x-webhook-secret` header

https://utari.up.railway.app/v1/triggers/[your-trigger-id]

curl -X POST 'https://utari.up.railway.app/v1/triggers/373a2...' \
  -H 'Content-Type: application/json' \
  -H 'x-webhook-secret: sk_trigger_So1ECq1Dat9hxcPip0OMPq912fvL9rPwKbu8bph3sW8' \
  -d '{"event": "example", "data": "your payload"}'

    Common locations:
    - Settings → Webhooks
    - Settings → Integrations → Webhooks
    - Settings → API → Webhooks
    - Developer → Webhooks
    - Automation → Webhooks

    1. Click "Add Webhook" or "Create Webhook"
    2. Paste Utari webhook URL as destination
    3. Select event types to trigger webhook
    4. Configure additional settings (if available)

    Examples:
    - Meeting completed
    - Form submitted
    - Payment received
    - File uploaded
    - Record created/updated    Select only events you want to process

    Some services request authentication:
    - Look for "Secret" or "Authentication" field
    - Paste the secret token from Utari
    - May be labeled as "Signing Secret" or "Verification Token"
    - Should be sent in x-webhook-secret header    Not all services support custom headers

    Common options:
    - Full payload (all data)
    - Transcript only
    - Metadata only
    - Custom fields    Choose based on your processing needs

    Save the webhook configuration    Webhook is now active!
    Events will trigger your Utari worker automatically

  Worker: "Meeting Transcription Processor"
  Integration: Fathom (with API key)
  Trigger: Webhook - Meeting completed

  Process this completed meeting recording.  The meeting data is: {{payload}}  Steps:
  1. Extract the transcript ID from the payload
  2. Use the Fathom integration to retrieve full transcript
  3. Analyze the transcript and extract:
     - Key discussion points
     - Action items and owners
     - Decisions made
     - Follow-up questions
  4. Create a meeting summary document
  5. Save to "Meeting Notes/[Date]/" folder
  6. Send summary to #meetings Slack channel  Execute automatically without asking for confirmation.

  Worker: "Lead Qualification Bot"
  Integration: CRM (Salesforce/HubSpot)
  Trigger: Webhook - Form submitted

  Process this form submission: {{payload}}  Extract:
  - Name: {{payload.name}}
  - Email: {{payload.email}}
  - Company: {{payload.company}}
  - Interest: {{payload.product_interest}}  Actions:
  1. Create new lead in CRM with extracted data
  2. Research company on LinkedIn for additional context
  3. Score lead based on:
     - Company size
     - Industry fit
     - Product interest
  4. If score > 7: Notify sales team via Slack immediately
  5. If score 4-7: Add to nurture sequence
  6. If score < 4: Add to general newsletter  Log all actions in "Lead Processing Log" spreadsheet.

  Worker: "Payment Processor"
  Integration: Email, Accounting system
  Trigger: Webhook - Payment succeeded

  Process successful payment: {{payload}}  Payment details:
  - Amount: {{payload.amount}}
  - Customer: {{payload.customer_email}}
  - Product: {{payload.product_name}}  Tasks:
  1. Send payment confirmation email to customer
  2. Log transaction in accounting spreadsheet
  3. If first payment: Send onboarding email sequence
  4. If renewal: Extend subscription, send thank you
  5. Notify finance team in Slack
  6. Update customer record in CRM  Execute immediately.

  Worker: "Meeting Prep Bot"
  Integration: Calendar, CRM, Email
  Trigger: Webhook - Event created with "Client Meeting" in title

  Prepare for upcoming meeting: {{payload}}  Meeting details:
  - Time: {{payload.start_time}}
  - Attendees: {{payload.attendees}}
  - Title: {{payload.title}}  Preparation:
  1. Extract company name from meeting title
  2. Research company and attendees on LinkedIn
  3. Check CRM for previous interactions
  4. Gather relevant product materials
  5. Create meeting prep document with:
     - Attendee backgrounds
     - Company context
     - Previous conversation summary
     - Suggested talking points
  6. Email prep document to meeting owner  Execute 24 hours before meeting.

    Go to: Worker → Triggers → Select webhook trigger    or    Go to: Triggers dashboard → Find webhook trigger

    See all webhook executions:
    - Timestamp
    - Status (success/failed)
    - Execution duration
    - Data received    Click on individual run to see details

    For each webhook execution:    View:
    - Prompt template used
    - Payload data received
    - Worker's response
    - Actions taken
    - Any errors    Use this to debug and improve prompts

✅ **Specific**:

    Process this meeting recording.
    Extract transcript ID from payload.transcript_id.
    Retrieve full transcript using Fathom integration.
    Analyze for action items and decisions.
    Create summary document.
    Send to Slack #meetings channel.

    I received meeting data. What should I do with it?

✅ **Direct execution**:

    Process automatically without asking for confirmation.
    Execute all steps immediately upon receiving webhook.

    Get the transcript and summarize it

✅ **Handles errors**:

    Get the transcript using the transcript_id.
    If transcript not available yet, wait 30 seconds and retry.
    If still unavailable after 3 attempts, log error and notify admin.
    If successful, proceed with summarization.

    Process the meeting

✅ **Uses specific fields**:

    Process meeting: {{payload.title}}
    Date: {{payload.date}}
    Duration: {{payload.duration}}
    Participants: {{payload.attendees}}    Use transcript_id: {{payload.transcript_id}} to retrieve full transcript