Skip to content

Run Agents Team

Orchestrates a team of AI agents to collaboratively solve a task using different coordination strategies.

Run Agents Team

Quick Start

  1. Create at least one agent using the Create Agent node.
  2. Connect one or more agents to the Run Agents Team node.
  3. Set the team type and provide a task description.
  4. Run the node to get the conversation and summary.

Setup Guide

1. Prepare Agents

  1. Use the Create Agent node to configure each agent.
  2. Connect the output of each agent to the corresponding agent input on this node.

2. Configure Team

  1. Select the team type (Cycle, Smart, or Manage).
  2. Enter the task for the agents to solve.
  3. Adjust summarization and message/timeout limits as needed.

Basic Usage

Multi-Agent Collaboration

  • Assign different roles or expertise to each agent.
  • Use for brainstorming, problem-solving, or multi-step tasks.
  • Choose the team type to control how agents interact.

Configuration

Required Inputs

Field Description Type Example
agent_1 First agent in the team. AGENT agent_1
team_type Coordination strategy: Cycle, Smart, or Manage. STRING Cycle
task Task description for the agents to solve. STRING "Summarize this text"
summarize Whether to output a summary of the conversation. BOOLEAN True
max_messages Maximum number of messages (0 = unlimited). INT 10
max_timeout Maximum execution time in seconds. INT 300

Optional Inputs

Field Description Type Example
agent_2 Second agent in the team. AGENT agent_2
agent_3 Third agent in the team. AGENT agent_3
... Up to 31 agents can be added in total. AGENT agent_31

Outputs

Field Description Example
messages Full conversation between the agents. Markdown transcript
summary Condensed summary of the agents' discussion. "Key findings..."

Best Practices

Team Composition

  • Assign clear, distinct roles to each agent for better collaboration.
  • Limit the number of agents to what is necessary for the task.

Task Design

  • Provide a clear, specific task description.
  • Use summarization for concise outputs when needed.

Troubleshooting

Common Issues

  • Agents do not respond: Ensure each agent is properly configured and connected.
  • Timeouts occur: Increase the max_timeout value for longer tasks.
  • No summary output: Check that the summarize option is enabled.

Need Help?

  • Review agent configurations and input values.
  • Check logs for error messages.
  • Consult platform documentation for further guidance.