This documentation is outdated. The platform is currently being upgraded to the latest version of ACP with major changes. Updated docs are expected in ~2 weeks. Apologies for the inconvenience.

​
GUI

The graphical user interface (GUI) provides an intuitive way to launch and manage agents.

Not all agents support GUI interaction.
1

Launch the web interface

beeai ui
2

Launch the agent

  • Navigate to the Agents tab.
  • Select you desired agent.
  • Click the Launch this agent button.

​
CLI

You can launch and interact with agents directly from your terminal using the command-line interface (CLI).

Basic syntax:

beeai run <agent_name>
To list all available agents, run: beeai list.

​
Interactive interface

Agents that implement a standardized interface support interactive mode.

Example usage:

# Chat agents
beeai run chat

# Hands-off agents
beeai run gpt-researcher

​
Simplified interface

Chat and hands-off agents support simplified invocation, accepting text part of an input as the first parameter:

# Chat agent example
beeai run chat "Hello"

# Hands-off agent example
beeai run gpt-researcher "Impact of climate change on global agriculture"

​
Complex inputs

For agents that require complex inputs or configurations, use JSON-formatted input:

# Agent requiring complex JSON input
beeai run content-judge '{
  "text": "Generate a concise summary of the history of artificial intelligence.",
  "agents": [
    "gpt-researcher",
    "ollama-deep-researcher"
  ]
}'

# Configuring a chat agent with the weather tool
beeai run chat '{
  "config": {
    "tools": ["weather"]
  },
  "messages": [{
    "role": "user",
    "content": "What is the weather in San Francisco?"
  }]
}'

​
ACP SDK

The ACP (Agent Communication Protocol) SDK lets you programmatically interact with agents.

Install it using:

uv add acp-sdk

Example of how to run an agent using ACP SDK:

import asyncio

from acp import ClientSession
from acp.client.sse import sse_client


async def run_client():
    async with sse_client(url="http://localhost:8333/mcp/sse") as streams:
        async with ClientSession(*streams) as session:
            await session.initialize()

            # Run the agent
            resp = await session.run_agent("gpt-researcher", {"text": "Impact of climate change on global agriculture"})
            print("Agent:", resp.output["text"])


asyncio.run(run_client())

​
BeeAI framework

The BeeAI Framework provides advanced tools to build and manage agent-based applications.

Install it via:

uv add beeai-framework

Example usage:

import asyncio
import json
from beeai_framework.agents.experimental.remote.agent import RemoteAgent


async def run():
    response = await RemoteAgent(
        url="http://localhost:8333/mcp/sse", agent_name="gpt-researcher"
    ).run("why is Nvidia stock going up?")

    print(json.loads(response.result.text))


def main():
    asyncio.run(run())

For more advanced examples, including partial updates and error logging, refer to the BeeAI Framework repository.