Skip to content

Three-Agent Collaboration

Script in langroid-examples

A full working example for the material in this section is in the three-agent-chat-num.py script in the langroid-examples repo: examples/quick-start/three-agent-chat-num.py.

Let us set up a simple numbers exercise between 3 agents. The Processor agent receives a number \(n\), and its goal is to apply a transformation to the it. However it does not know how to apply the transformation, and takes the help of two other agents to do so. Given a number \(n\),

  • The EvenHandler returns \(n/2\) if n is even, otherwise says DO-NOT-KNOW.
  • The OddHandler returns \(3n+1\) if n is odd, otherwise says DO-NOT-KNOW.

We'll first define a shared LLM config:

llm_config = lr.language_models.OpenAIGPTConfig(
    chat_model=lr.language_models.OpenAIChatModel.GPT4o,
    # or, e.g., "ollama/qwen2.5-coder:latest", or "gemini/gemini-2.0-flash-exp"
)

Next define the config for the Processor agent: 

processor_config = lr.ChatAgentConfig(
    name="Processor",
    llm = llm_config,
    system_message="""
    You will receive a number from the user.
    Simply repeat that number, DO NOT SAY ANYTHING else,
    and wait for a TRANSFORMATION of the number 
    to be returned to you.

    Once you have received the RESULT, simply say "DONE",
    do not say anything else.
    """,        
    vecdb=None,
)

Then set up the processor_agent, along with the corresponding task: 

processor_agent = lr.ChatAgent(processor_config)

processor_task = lr.Task(
    processor_agent,
    llm_delegate=True, #(1)!
    interactive=False, #(2)!
    single_round=False, #(3)!
)

  1. Setting the llm_delegate option to True means that the processor_task is delegated to the LLM (as opposed to the User), in the sense that the LLM is the one "seeking" a response to the latest number. Specifically, this means that in the processor_task.step() when a sub-task returns DO-NOT-KNOW, it is not considered a valid response, and the search for a valid response continues to the next sub-task if any.
  2. interactive=False means the task loop will not wait for user input.
  3. single_round=False means that the processor_task should not terminate after a valid response from a responder.

Set up the other two agents and tasks:

NO_ANSWER = lr.utils.constants.NO_ANSWER

even_config = lr.ChatAgentConfig(
    name="EvenHandler",
    llm = llm_config,
    system_message=f"""
    You will be given a number N. Respond as follows:

    - If N is even, divide N by 2 and show the result, 
      in the format: 
        RESULT = <result>
      and say NOTHING ELSE.
    - If N is odd, say {NO_ANSWER}
    """,    
)
even_agent = lr.ChatAgent(even_config)
even_task = lr.Task(
    even_agent,
    single_round=True,  # task done after 1 step() with valid response
)

odd_config = lr.ChatAgentConfig(
    name="OddHandler",
    llm = llm_config,
    system_message=f"""
    You will be given a number N. Respond as follows:

    - if N is odd, return the result (N*3+1), in the format:
        RESULT = <result> 
        and say NOTHING ELSE.

    - If N is even, say {NO_ANSWER}
    """,
)
odd_agent = lr.ChatAgent(odd_config)
odd_task = lr.Task(
    odd_agent,
    single_round=True,  # task done after 1 step() with valid response
)

Now add the even_task and odd_task as subtasks of the processor_task, and then run it with a number as input:

processor_task.add_sub_task([even_task, odd_task])
processor_task.run(13)

The input number will be passed to the Processor agent as the user input.

Feel free to try the working example script three-agent-chat-num.py langroid-examples repo: examples/quick-start/three-agent-chat-num.py:

python3 examples/quick-start/three-agent-chat-num.py

Here's a screenshot of what it looks like: three-agent-num.png

Next steps

In the next section you will learn how to use Langroid to equip a ChatAgent with tools or function-calling.