Skip to content

Agno Framework

Agno coordinates and orchestrates role-playing autonomous AI agents. Each agent has:

  • Name and Role: what they do
  • Goal: their objective
  • Backstory: context
  • Tools: available abilities

Tasks are assigned to agents with expected outputs.

Example

multi-agent-generator "Research AI trends and write a summary" --framework agno

Produces agents like:

{
  "model_id": "gpt-4o",
  "process": "sequential",
  "agents": [
    {
      "name": "research_specialist",
      "role": "Research Specialist",
      "goal": "Gather AI research trends",
      "backstory": "Expert in sourcing and aggregating technology news",
      "tools": ["DuckDuckGoTools", "Newspaper4kTools"]
    },
    {
      "name": "writer",
      "role": "Content Writer",
      "goal": "Write a clear summary",
      "backstory": "Skilled at concise technical writing",
      "tools": []
    }
  ],
  "tasks": [
    {
      "name": "research_task",
      "description": "Find recent AI trends across news and blogs",
      "agent": "research_specialist",
      "expected_output": "Bullet list of trends with links"
    },
    {
      "name": "writing_task",
      "description": "Summarize the trends for a general audience",
      "agent": "writer",
      "expected_output": "400-word Markdown summary"
    }
  ]
}