The following tools are exposed by superglue’s MCP server. The input schemas are defined using Zod in mcp-server.ts.

​
Workflow Discovery & Management

​
superglue_list_available_workflows

Lists all available superglue workflows for the current organization. Input Schema:
  • limit: (Optional) Number of workflows to return (default: 100)
  • offset: (Optional) Offset for pagination (default: 0)
Returns:
  • List of workflows with their IDs, instructions, timestamps, and saved credentials
  • Use the workflow IDs with superglue_execute_workflow to run specific workflows
Example Usage: 
{
  "toolName": "superglue_list_available_workflows",
  "inputs": {
    "limit": 50,
    "offset": 0
  }
}

​
superglue_find_relevant_integrations

Finds integrations relevant to a given natural language instruction. Used as a first step before building a new workflow. Input Schema:
  • instruction: (Optional) Natural language description of what you want to do. If not provided, returns all available integrations.
Returns:
  • List of suggested integration IDs with reasons and available credentials
  • If no integrations exist, returns empty list and sugggests creating new integrations
Example Usage: 
{
  "toolName": "superglue_find_relevant_integrations",
  "inputs": {
    "instruction": "I need to sync data between Stripe and HubSpot"
  }
}

​
Workflow Execution

​
superglue_execute_workflow

Executes a previously saved superglue workflow by its ID. Input Schema:
  • id: Required - The ID of the workflow to execute
  • payload: (Optional) JSON payload to pass to the workflow
  • credentials: (Optional) Additional credentials that will be merged with integration credentials
  • options: (Optional) Request configuration (caching, timeouts, retries, etc.)
Important Notes:
  • This tool is for running existing, saved workflows only
  • To create a new workflow, use superglue_build_and_run
  • Workflow ID must exist (use superglue_list_available_workflows to find valid IDs)
Example Usage: 
{
  "toolName": "superglue_execute_workflow",
  "inputs": {
    "id": "stripe-to-hubspot-sync",
    "payload": { "customerId": "cus_123" },
    "credentials": { 
      "additionalApiKey": "sk_test_..."
    }
  }
}

​
Workflow Building & Testing

​
superglue_build_and_run

Builds and executes workflows. This is the primary tool for creating and iteratively testing workflows. Input Schema:
  • instruction: Required - Natural language instruction to build a new workflow from scratch
  • integrationIds: Required - Array of integration IDs to use in the workflow
  • payload: (Optional) JSON payload for the workflow execution
  • credentials: (Optional) Additional credentials that will be merged with integration credentials
  • responseSchema: (Optional) JSONSchema for the expected output structure
Important Notes:
  • This tool only builds and tests workflows - it does NOT save them
  • Building and testing can take up to 1 minute
  • Use superglue_find_relevant_integrations first to discover available integration IDs
  • After successful execution, use superglue_save_workflow to persist the workflow
Example Usage: 
{
  "toolName": "superglue_build_and_run",
  "inputs": {
    "instruction": "Fetch customer data from Stripe and create or update contact in HubSpot",
    "integrationIds": ["stripe", "hubspot"],
    "payload": { "customerId": "cus_123" },
    "responseSchema": {
      "type": "object",
      "properties": {
        "hubspotContactId": { "type": "string" },
        "status": { "type": "string" }
      }
    }
  }
}

​
superglue_save_workflow

Saves a previously built and tested workflow. Use this after successful execution of superglue_build_and_run. Input Schema:
  • id: Required - Unique identifier for the workflow to save
  • workflow: Required - Workflow configuration object from build_and_run result
Important Notes:
  • Take the workflow data from build_and_run result’s config field
  • DO NOT set any fields to null - omit optional fields entirely
  • Each step MUST have an integrationId field
  • Workflow MUST have integrationIds array
Example Usage: 
{
  "toolName": "superglue_save_workflow",
  "inputs": {
    "id": "stripe-to-hubspot-sync",
    "workflow": {
      "steps": [...],
      "integrationIds": ["stripe", "hubspot"],
      "instruction": "Sync Stripe customers to HubSpot",
      "finalTransform": "$",
      "responseSchema": {...}
    }
  }
}

​
Integration Management

​
superglue_create_integration

Creates and immediately saves a new integration. Integrations are building blocks for workflows and contain the credentials for accessing APIs. Input Schema:
  • id: Required - A unique identifier for the new integration
  • name: (Optional) Human-readable name for the integration
  • urlHost: (Optional) Base URL/hostname for the API including protocol
  • urlPath: (Optional) Path component of the URL
  • documentationUrl: (Optional) URL to the API documentation
  • documentation: (Optional) API documentation content, if provided directly
  • credentials: Required - Credentials object (can be empty if no credentials needed)
Important Notes:
  • Most APIs require authentication (API keys, tokens, etc.)
  • Always store credentials in the credentials field
  • Use placeholder references: <<{integration_id}_{credential_name}>>
  • Split information clearly: urlHost (without secrets), credentials (with secrets)
  • Providing a documentationUrl triggers async documentation processing
Example Usage: 
{
  "toolName": "superglue_create_integration",
  "inputs": {
    "id": "my-api",
    "name": "My Custom API",
    "urlHost": "https://api.example.com",
    "credentials": {
      "apiKey": "sk_live_abc123"
    },
    "documentationUrl": "https://api.example.com/docs"
  }
}

​
Code Generation

​
superglue_get_workflow_integration_code

Generate integration code for a specific workflow. Use this to show users how to implement a workflow in their applications. Input Schema:
  • workflowId: Required - The ID of the workflow to generate code for
  • language: Required - Programming language: typescript, python, or go
Returns:
  • Ready-to-use SDK code for the specified language
  • Includes example payload and credentials based on the workflow’s input schema
Example Usage: 
{
  "toolName": "superglue_get_workflow_integration_code",
  "inputs": {
    "workflowId": "stripe-to-hubspot-sync",
    "language": "typescript"
  }
}

​
Agent Workflow

The recommended workflow for agents using the Superglue MCP server:
  1. DISCOVER: Use superglue_find_relevant_integrations to find available integrations for your task
  2. BUILD & TEST: Use superglue_build_and_run with instruction and integrations. Iterate until successful
  3. SAVE (Optional): Ask user if they want to save the workflow, then use superglue_save_workflow with the workflow data
  4. EXECUTE: Use superglue_execute_workflow for saved workflows
  5. INTEGRATE: Use superglue_get_workflow_integration_code to generate SDK code
Best Practices
  • Always start with superglue_find_relevant_integrations for discovery
  • Create integrations and store credentials using superglue_create_integration
  • Ask users for credentials if needed
  • Ask user before saving workflows
  • When saving workflows, NEVER set fields to null - omit optional fields if no value available
  • Copy actual values from build_and_run results, don’t assume fields are empty