Skip to content

Multi-Agent Guide

Kabot can run multiple agents with separate workspaces, models, and routing rules.

Why Use Multiple Agents

Use multiple agents when one bot should not do everything.

Examples: - work agent - personal agent - support agent - research agent - automation agent

What Stays Separate Per Agent

  • workspace
  • session history
  • model path
  • bindings and usage context

Simple Example

agents:
  list:
    - id: work
      name: Work Agent
      model: openai/gpt-4o
      workspace: ~/.kabot/workspace-work
    - id: personal
      name: Personal Agent
      model: anthropic/claude-sonnet-4-5
      workspace: ~/.kabot/workspace-personal
      default: true

Routing Basics

Messages generally resolve by: 1. exact binding 2. channel binding 3. default agent fallback

Best Practices

  • give each agent a clear role
  • keep one default agent
  • use separate workspaces
  • avoid overloading one agent with every channel and every task

When Multi-Agent Is Worth It

Choose multi-agent when: - different channels need different models - different users need different personas or boundaries - you want cleaner memory separation - you need dedicated role-based automation