Skip to content

MajorityVoting: 3-Step Quickstart Guide

The MajorityVoting architecture enables multiple agents to independently evaluate a task, then synthesizes their responses through a consensus agent that evaluates, ranks, and provides a comprehensive final answer. This creates more robust and well-rounded solutions through diverse perspectives.

Overview

Feature Description
Concurrent Execution Multiple agents run simultaneously for faster processing
Consensus Building Dedicated consensus agent evaluates and synthesizes all responses
Multi-Loop Support Iterative refinement over multiple rounds with agent memory
Comprehensive Evaluation Assesses accuracy, depth, relevance, clarity, and unique insights 
Agent A ─┐
Agent B ─┼──> Concurrent Execution
Agent C ─┘
    │
    ▼
 Consensus Agent
    │
    ▼
Synthesized Final Answer

Step 1: Install and Import

Ensure you have Swarms installed and import the MajorityVoting class:

pip install swarms

from swarms import Agent, MajorityVoting

Step 2: Create the Voting System

Create agents and initialize the MajorityVoting system:

# Create specialized agents
agent1 = Agent(
    agent_name="Financial-Analysis-Agent-1",
    system_prompt="You are a conservative financial advisor focused on risk management and long-term stability.",
    model_name="gpt-4o-mini",
    max_loops=1,
)

agent2 = Agent(
    agent_name="Financial-Analysis-Agent-2",
    system_prompt="You are a growth-oriented financial advisor focused on high-potential opportunities.",
    model_name="gpt-4o-mini",
    max_loops=1,
)

agent3 = Agent(
    agent_name="Financial-Analysis-Agent-3",
    system_prompt="You are a balanced financial advisor focused on diversification and risk-adjusted returns.",
    model_name="gpt-4o-mini",
    max_loops=1,
)

# Create the majority voting system
voting_system = MajorityVoting(
    agents=[agent1, agent2, agent3],
    max_loops=1,
    verbose=True
)

Step 3: Run the Voting System

Execute a task and get the consensus result:

# Define the task
task = "Create a table of super high growth opportunities for AI. I have $40k to invest in ETFs, index funds, and more. Please create a table in markdown."

# Run the voting system
result = voting_system.run(task=task)

# Print the result
print(result)

Complete Example

Here's a complete working example:

from swarms import Agent, MajorityVoting

# Step 1: Create specialized agents with different perspectives
agents = [
    Agent(
        agent_name="Conservative-Analyst",
        system_prompt="You are a conservative financial advisor focused on risk management and long-term stability.",
        model_name="gpt-4o-mini",
        max_loops=1,
    ),
    Agent(
        agent_name="Growth-Analyst",
        system_prompt="You are a growth-oriented financial advisor focused on high-potential opportunities.",
        model_name="gpt-4o-mini",
        max_loops=1,
    ),
    Agent(
        agent_name="Balanced-Analyst",
        system_prompt="You are a balanced financial advisor focused on diversification and risk-adjusted returns.",
        model_name="gpt-4o-mini",
        max_loops=1,
    ),
]

# Step 2: Create the majority voting system
voting_system = MajorityVoting(
    agents=agents,
    max_loops=1,  # Single round of voting
    output_type="dict",  # Return as dictionary
    verbose=True
)

# Step 3: Run a task
task = "What are the top 3 AI investment opportunities for 2024?"
result = voting_system.run(task=task)

print("=" * 60)
print("VOTING RESULT:")
print("=" * 60)
print(result)

Multi-Loop Consensus Building

Enable iterative refinement with multiple rounds:

from swarms import Agent, MajorityVoting

# Create agents
agents = [
    Agent(agent_name="Analyst-1", system_prompt="Expert financial analyst", model_name="gpt-4o-mini", max_loops=1),
    Agent(agent_name="Analyst-2", system_prompt="Expert market researcher", model_name="gpt-4o-mini", max_loops=1),
    Agent(agent_name="Analyst-3", system_prompt="Expert risk assessor", model_name="gpt-4o-mini", max_loops=1),
]

# Multi-loop voting system
voting_system = MajorityVoting(
    agents=agents,
    max_loops=3,  # 3 rounds of iterative refinement
    verbose=True
)

# Each loop:
# 1. Agents independently analyze the task
# 2. Consensus agent evaluates and synthesizes responses
# 3. Agents see the consensus and refine their responses in next loop

result = voting_system.run("Analyze the future of quantum computing in enterprise applications")
print(result)

Batch Processing

Process multiple tasks sequentially:

from swarms import Agent, MajorityVoting

voting_system = MajorityVoting(
    agents=[agent1, agent2, agent3],
    max_loops=1
)

# Multiple tasks
tasks = [
    "What are the best cloud computing stocks?",
    "Should I invest in renewable energy ETFs?",
    "What's the outlook for semiconductor companies?",
]

# Process all tasks
results = voting_system.batch_run(tasks)

for task, result in zip(tasks, results):
    print(f"\nTask: {task}")
    print(f"Result: {result[:200]}...")

Configuration Options

Parameter Default Description
agents Required List of Agent instances to participate in voting
max_loops 1 Number of consensus rounds for iterative refinement
output_type "dict" Output format: "dict", "str", "list"
verbose False Enable detailed logging
consensus_agent_model_name "gpt-4.1" Model for the consensus agent
consensus_agent_prompt Default Custom system prompt for consensus agent

Output Types

Value Description
"dict" Conversation history as dictionary with roles and content
"str" All messages formatted as a single string
"list" Messages as a list of dictionaries

Use Cases

Domain Example Task
Financial Analysis "Compare investment strategies for retirement planning"
Technical Decisions "Evaluate database options for a high-traffic application"
Research "Synthesize findings on climate change mitigation strategies"
Product Development "Analyze feature prioritization for our mobile app"
Strategic Planning "Evaluate market entry strategies for Southeast Asia"

How It Works

  1. Concurrent Execution: All agents run simultaneously on the same task
  2. Independent Analysis: Each agent provides its own perspective based on its system prompt
  3. Consensus Evaluation: The consensus agent:
  4. Evaluates each agent's response on multiple dimensions
  5. Compares and contrasts different viewpoints
  6. Identifies the strongest arguments
  7. Synthesizes a comprehensive final answer
  8. Iterative Refinement (if max_loops > 1): Agents see the consensus and refine their responses

Best Practices

  • Diverse Perspectives: Create agents with different specializations or viewpoints
  • Clear Prompts: Give each agent a distinct role and expertise area
  • Appropriate Loops: Use max_loops=1 for simple consensus, higher for complex iterative refinement
  • Output Format: Use output_type="dict" to inspect individual agent responses
  • Consensus Quality: The default consensus prompt is optimized for comprehensive evaluation

Next Steps