Overview

The AIOS-Agent SDK (Cerebrum) provides a structured interface between user device applications and the AIOS kernel through different functional modules.

How agents leverage Agent-SDK to interact with AIOS kernel

All queries from these modules are ultimately channeled through a central send_request(Query) function in the AIOS-Agent SDK, which then communicates with the AIOS kernel via HTTP requests (either to localhost or a remote URL).

On the user device side, the agent contains code snippets for planning, action, memory, and storage that interact with their respective modules through the Cerebrum Client interface. This creates a clean separation between the application logic and the kernel communication layer.

To use the SDK to interact with the AIOS kernel, it is required to set up a client like the following

Step 1: Initialize Your Client

First, let's create your client instance:

from cerebrum import config
from cerebrum.client import Cerebrum

client = Cerebrum()
config.global_client = client

Step 2: Add Functionality Layers

The AIOS-Agent SDK offers five router APIs to initialize and setup the following essential modules in the AIOS kernel.

• LLM

• Memory

• Storage

• Tools

• Scheduler

Here's a simple example of how to add these layers to your client, and how to set up the LLMLayer, MemoryLayer, StorageLayer and ToolLayer will be introduced in the later sections.

from cerebrum.llm.layer import LLMLayer
from cerebrum.memory.layer import MemoryLayer
from cerebrum.overrides.layer import OverridesLayer
from cerebrum.storage.layer import StorageLayer
from cerebrum.tool.layer import ToolLayer

client.add_llm_layer(
    LLMLayer(llm_name=llm_name, llm_backend=llm_backend)  # Configure your LLM
).add_storage_layer(
    StorageLayer(root_dir=root_dir)  # Set storage directory
).add_memory_layer(
    MemoryLayer(memory_limit=memory_limit)  # Set memory per agent
).add_tool_layer(
    ToolLayer()  # Add tool capabilities
).override_scheduler(
    OverridesLayer(max_workers=max_workers)  # Configure scheduling
)

Step 3: Run Your Agent

Now you can run agents and get their results:

try:
    # Connect to the client
    client.connect()
    
    # Execute your agent
    agent_path = ""  # Your agent's name or path
    task = ""       # Your task description
    result = client.execute(agent_path, {"task": task})
    
    # Get the results
    final_result = client.poll_agent(
        result["execution_id"],
        timeout=timeout
    )
    return final_result

except TimeoutError:
    return None
except Exception as e:
    return None
finally:
    client.cleanup()