Skip to main content

Master Agent

Orchestrate and coordinate multiple agent nodes to execute complex AI workflows.

What is the Master Agent?​

The Master Agent is an advanced orchestration node that coordinates multiple specialized agents to execute complex AI workflows. It acts as the central controller that manages task delegation and workflow execution through the agents you connect to it.

How to use it?​

To set up and configure the Master Agent for your workflows:

  1. Select Language Model:

    • Choose from advanced models like OpenAI GPT-5, GPT-4o, Claude, or AWS Bedrock models
    • Consider model capabilities for complex reasoning and coordination tasks
    • Select appropriate credentials and region for your chosen model

  2. Configure Model Parameters:

    • Temperature: Set creativity level (0.0-1.0, default: 0.7) - lower values for more consistent coordination
    • Top P: Control response diversity (0.0-1.0, default: 1.0)
    • Response Length: Define maximum response length based on your model's limits
    • Maximum Steps: Set the maximum number of reasoning steps the agent can take

  3. Provide Your Task:

    • Prompt: Clearly describe your objective and what you want to accomplish
    • Be specific about your requirements and expected outcomes
    • Include any constraints or special considerations for the workflow

  4. Connect Specialized Agents (Optional):

    • Planning Agent: Connect to break down complex tasks into actionable steps
    • Executor Agent: Connect to implement and execute the planned actions
    • Validation Agent: Connect to analyze results and provide quality improvements

  5. Select Available Tools:

    • Choose from your available tools that the Master Agent can use
    • Select tools relevant to your specific workflow requirements
    • The agent will intelligently use these tools as needed during execution

Example of usage​

Objective: Create a research and content creation workflow using the Master Agent with connected specialized agents.

Master Agent Setup:

  1. Model Configuration:

    • Select OpenAI GPT-4o or Claude 3.7 Sonnet for strong reasoning capabilities
    • Set temperature to 0.3 for consistent coordination decisions
    • Configure response length to 4096 tokens for detailed planning

  2. Agent Connections:

    • Connect a Planning Agent to break down the research requirements systematically
    • Connect an Executor Agent to implement data collection and content creation steps
    • Connect a Validation Agent to review and improve the generated content

  3. Tool Selection:

    • Select File Reader for processing research documents
    • Add Website Scraper for gathering online information
    • Include Text Embedder and Vector Store for knowledge management
    • Add LLM nodes for content generation tasks

  4. Task Input:

    • Provide clear prompt: "Research topic X and create a comprehensive report with citations"
    • Specify requirements like word count, format, and quality standards

Expected Workflow:

  • Master Agent coordinates with Planning Agent to create research strategy
  • Executor Agent implements data collection using selected tools
  • Validation Agent reviews outputs and suggests improvements
  • Master Agent synthesizes all results into final deliverable

Additional information​

Configuration Options:​

Model Selection:

  • Choose models based on your complexity requirements and budget
  • GPT-5 and GPT-4o: Best for complex reasoning and coordination
  • Claude models: Excellent for analytical tasks and detailed planning
  • Llama models: Cost-effective options for simpler coordination tasks

Temperature Settings:

  • Low (0.1-0.3): For consistent, predictable coordination decisions
  • Medium (0.4-0.7): Balanced approach with some creative problem-solving
  • High (0.8-1.0): For creative workflows requiring innovative approaches

Agent Connections:

  • Planning Agent: Essential for breaking down complex multi-step tasks
  • Executor Agent: Required when you need actual implementation of planned steps
  • Validation Agent: Important for quality assurance and improvement cycles
  • All Three: Recommended for comprehensive, high-quality workflows. If some of the agents are not configured the Master Agent creates its own agents to take over the tasks.

Setup Best Practices:​

Task Definition:

  • Write clear, specific prompts describing your exact requirements
  • Include success criteria and any constraints or preferences
  • Provide context about the intended use of the output

Tool Selection:

  • Choose tools that match your workflow requirements
  • Include data sources (File Reader, Website Scraper) for information gathering
  • Add processing tools (Text Embedder, Document Splitter) for data handling
  • Select output tools (File Writer, Email) for delivering results

Model Configuration:

  • Start with moderate temperature (0.5-0.7) and adjust based on results
  • Set appropriate response length limits for your use case
  • Use step limits to prevent excessive processing time

Common Use Cases:​

Research and Analysis:

  • Literature reviews with multiple data sources
  • Market research combining web scraping and analysis
  • Technical documentation creation with fact-checking

Content Creation:

  • Blog posts with research, writing, and quality review phases
  • Marketing materials requiring planning, creation, and optimization
  • Educational content with structured approach and validation

Data Processing:

  • Document analysis workflows with processing and summarization
  • Information extraction from multiple file types and sources
  • Knowledge base creation with organization and quality control

Automated Reporting:

  • Regular reports combining data collection and analysis
  • Status updates pulling information from various sources
  • Compliance reports requiring systematic data gathering and validation

Input and Output:​

What You Provide:

  • Prompt: Clear description of your task and requirements
  • Connected Agents: Optional specialized agents for complex workflows
  • Available Tools: Selection of tools the Master Agent can use
  • Configuration: Model settings, temperature, and response limits

What You Receive:

  • Final Result: Completed output based on your prompt requirements
  • Coordinated Workflow: Results from orchestrated multi-agent collaboration
  • Quality Assured Output: Results improved through validation cycles

The Master Agent serves as your central coordinator, intelligently managing whatever agents and tools you provide to accomplish your specified objectives.