Skip to main content

Model Context Protocol (MCP)

Early Access

Try our newest feature! Only available through reach out. Contact us to join the early access program and get priority support.

Unizo offers an MCP server that integrates incident management platforms with any LLM provider supporting the MCP protocol. This enables your AI agent to create incidents, coordinate responses, and analyze incident patterns across PagerDuty, Opsgenie, Incident.io, and other incident management tools via a single MCP server.

Supported Tools & Use Cases

The following tools are available in the Incident management MCP Server:

Tool NameDescription
list_incidentsBrowse and search active and historical incidents with filtering
create_incidentDeclare new incidents with proper categorization and routing
update_incidentUpdate incident status, severity, or other properties
get_oncallFind on-call responders for services and escalation levels
add_respondersPage additional team members to join incident response
post_updateShare incident updates, timelines, and status communications

list_incidents

Browse and search active and historical incidents with filtering

Parameters:

  • integration: Target incident management platform (required)
  • status: Filter by incident status (triggered, acknowledged, resolved)
  • severity: Filter by severity (P1, P2, P3, P4)
  • service: Filter by affected service
  • time_range: Time period to search
  • assignee: Filter by assigned responder
  • page: Pagination control
  • limit: Number of results per page

create_incident

Declare new incidents with proper categorization and routing

Parameters:

  • integration: Target incident management platform (required)
  • title: Incident title (required)
  • description: Detailed incident description (required)
  • severity: Incident severity (P1-P4)
  • service: Affected service or component
  • tags: Labels for categorization
  • custom_fields: Platform-specific fields
  • urgency: Response urgency level

update_incident

Update incident status, severity, or other properties

Parameters:

  • integration: Target incident management platform (required)
  • incident_id: Incident to update (required)
  • status: New incident status
  • severity: Update severity level
  • resolution: Resolution details
  • impact: Current impact assessment
  • custom_fields: Update custom field values

get_oncall

Find on-call responders for services and escalation levels

Parameters:

  • integration: Target incident management platform (required)
  • services: Services to check on-call for (required)
  • escalation_level: Which escalation level to check
  • time: Specific time to check (default: now)
  • include_escalation: Include full escalation policy

add_responders

Page additional team members to join incident response

Parameters:

  • integration: Target incident management platform (required)
  • incident_id: Target incident (required)
  • responders: List of responders to add (required)
  • message: Custom page message
  • urgency: Override urgency for page
  • escalation_level: Target specific escalation level

post_update

Share incident updates, timelines, and status communications

Parameters:

  • integration: Target incident management platform (required)
  • incident_id: Target incident (required)
  • message: Update message (required)
  • update_type: Type of update (investigation, status, resolution)
  • visibility: Public or internal update
  • notify_subscribers: Send notifications

Installation

Prerequisites

  • A Unizo API key
  • An active Incidents integration (ServiceNow, Jira Service Management, ManageEngine ServiceDesk Plus, Opsgenie, PagerDuty, SolarWinds, VictorOps)
  • Node.js v20 or higher

MCP Configuration

Example configuration for running the Unizo Incident Management MCP server:

{
"mcpServers": {
"unizo-incidents": {
"command": "npx",
"args": [
"mcp-remote",
"http://api.unizo.ai/mcp/incidents",
"--allow-http",
"--header",
"apikey:${UNIZO_API_KEY}"
],
"env": {
"UNIZO_API_KEY": "your_api_key"
}
}
}
}

Client Setup

For detailed setup instructions with specific AI clients:

Environment Variables

The following environment variables are required:

Error Handling

All tools return errors in a consistent format:

{
"error": {
"code": "INCIDENT_NOT_FOUND",
"message": "Incident 'abc-123' not found"
}
}

Common error codes:

  • INTEGRATION_NOT_FOUND: Invalid incident management integration ID
  • INCIDENT_NOT_FOUND: Specified incident does not exist
  • RESPONDER_NOT_FOUND: Responder not found or not assignable
  • UNAUTHORIZED: Invalid API key or insufficient permissions
  • RATE_LIMIT_EXCEEDED: API quota exceeded
  • INVALID_STATUS: Status update value not accepted by platform