Huawei Cloud AgentArts SDK

Build, deploy and manage AI agents with Huawei Cloud capabilities.

Overview

Huawei Cloud AgentArts SDK is a comprehensive toolkit for developing, deploying, and managing AI agents. It provides seamless integration with Huawei Cloud services while supporting mainstream agent frameworks.

Key Features

• Framework Agnostic - Compatible with LangChain, LangGraph, AutoGen, CrewAI, Google ADK, and any custom agent framework
• One-Click Deployment - Deploy agents to Huawei Cloud with a single command
• Built-in Tools - Code interpreter sandbox, memory management, gateway support
• Cloud Integration - Seamless integration with Huawei Cloud authentication, monitoring, and logging
• CLI Toolkit - Complete command-line tools for project initialization, local development, and cloud deployment

Repository Structure

agentarts-sdk-python/
├── src/agentarts/
│   ├── sdk/                    # Core SDK modules
│   │   ├── runtime/            # HTTP server runtime (AgentArtsRuntimeApp)
│   │   ├── memory/             # Conversation memory management
│   │   ├── tools/              # Built-in tools (Code Interpreter)
│   │   ├── gateway/            # Gateway client
│   │   ├── identity/           # Authentication & authorization
│   │   ├── integration/        # Framework adapters (LangGraph, etc.)
│   │   └── service/            # HTTP clients for cloud services
│   └── toolkit/                # CLI toolkit
│       ├── cli/                # Command-line interface
│       ├── operations/         # CLI operation handlers
│       ├── plugins/memory/    # Memory plugins
│       │   ├── ai_agent/      # Platform assets (claude_code, codex, opencode, hermes)
│       │   ├── installer/     # Unified installer
│       │   ├── mcp/           # MCP server (local adapter)
│       │   └── resources/     # Manifests + shared hook scripts
│       └── utils/             # Utilities
│           ├── runtime/       # Runtime utilities
│           └── templates/     # Project templates
├── docs/                       # Documentation
│   └── cn/                     # Chinese documentation
│       ├── sdk_user_guide/     # SDK usage guides
│       └── toolkit_user_guide/ # CLI usage guides
├── examples/                   # Example projects
└── tests/                      # Test suites
    ├── unit/                   # Unit tests mirroring src/ tree
    │   ├── sdk/                # SDK tests
    │   └── toolkit/            # Toolkit tests
    │       └── plugins/memory/ # Memory plugin tests
    └── integration/            # Integration tests

Wrapping Your Agent as HTTP Server

The SDK provides AgentArtsRuntimeApp to wrap your agent logic as a standard HTTP server, exposing:

• POST /invocations - Main agent invocation endpoint
• GET /ping - Health check endpoint
• WS /ws - WebSocket endpoint for streaming

Example: LangGraph Agent

# agent.py
import os
from typing import Dict, Any, TypedDict, Annotated
from operator import add

from agentarts.sdk import AgentArtsRuntimeApp, RequestContext

app = AgentArtsRuntimeApp()


class State(TypedDict):
    messages: Annotated[list, add]
    query: str
    response: str


class LangGraphAgent:
    def __init__(self):
        self.model_name = os.environ.get("OPENAI_MODEL_NAME", "gpt-4o-mini")
        self._graph = None

    def _build_graph(self):
        from langgraph.graph import StateGraph, END
        from langchain_openai import ChatOpenAI
        from langchain_core.messages import HumanMessage, AIMessage

        llm = ChatOpenAI(
            model=self.model_name,
            api_key=os.environ.get("OPENAI_API_KEY"),
            base_url=os.environ.get("OPENAI_BASE_URL")
        )

        async def process_node(state: State) -> Dict[str, Any]:
            query = state.get("query", "")
            messages = state.get("messages", []) or [HumanMessage(content=query)]
            response = await llm.ainvoke(messages)
            return {
                "messages": [AIMessage(content=response.content)],
                "response": response.content,
            }

        workflow = StateGraph(State)
        workflow.add_node("process", process_node)
        workflow.set_entry_point("process")
        workflow.add_edge("process", END)
        return workflow.compile()

    async def run(self, query: str) -> Dict[str, Any]:
        graph = self._graph or self._build_graph()
        self._graph = graph
        result = await graph.ainvoke({"messages": [], "query": query, "response": ""})
        return {"response": result.get("response", "")}


_agent = LangGraphAgent()


@app.entrypoint
async def handler(payload: Dict[str, Any], context: RequestContext = None) -> Dict[str, Any]:
    query = payload.get("message", "")
    return await _agent.run(query)


if __name__ == "__main__":
    app.run()

Key Points

1. Focus on Agent Logic - You only need to implement the agent logic; the SDK handles HTTP server, request parsing, and response formatting
2. Framework Agnostic - Works with any agent framework (LangChain, LangGraph, AutoGen, CrewAI, or custom implementations)
3. Simple Decorator - Use @app.entrypoint to mark your handler function
4. Context Support - Optional RequestContext parameter provides session info and request metadata
5. Configurable Model - Model name can be configured via environment variable (e.g., OPENAI_MODEL_NAME)

Installation

Requirements

• Python 3.10 or higher
• pip or uv package manager

Create Virtual Environment (Recommended)

It is recommended to install the SDK in a virtual environment to avoid dependency conflicts.

Windows:

# Create virtual environment
python -m venv venv

# Activate virtual environment
.\venv\Scripts\Activate.ps1

# Or using Command Prompt
.\venv\Scripts\activate.bat

Linux/macOS:

# Create virtual environment
python -m venv venv

# Activate virtual environment
source venv/bin/activate

Install via pip

Windows:

pip install agentarts-sdk

Linux/macOS:

pip install agentarts-sdk

Install with Optional Dependencies

# With LangChain support
pip install agentarts-sdk[langchain]

# With LangGraph support
pip install agentarts-sdk[langgraph]

# With MCP memory server support (required by `agentarts memory install`)
pip install agentarts-sdk[mcp]

# With all optional dependencies
pip install agentarts-sdk[all]

Install from Source

Windows:

git clone https://github.com/huaweicloud/agentarts-sdk-python.git
cd agentarts-sdk-python

# Create and activate virtual environment
python -m venv venv
.\venv\Scripts\Activate.ps1

# Install in development mode
pip install -e ".[dev]"

Linux/macOS:

git clone https://github.com/huaweicloud/agentarts-sdk-python.git
cd agentarts-sdk-python

# Create and activate virtual environment
python -m venv venv
source venv/bin/activate

# Install in development mode
pip install -e ".[dev]"

Configure Huawei Cloud Credentials

Set environment variables for Huawei Cloud authentication:

Windows (PowerShell):

$env:HUAWEICLOUD_SDK_AK = "your-access-key"
$env:HUAWEICLOUD_SDK_SK = "your-secret-key"

Windows (Command Prompt):

set HUAWEICLOUD_SDK_AK=your-access-key
set HUAWEICLOUD_SDK_SK=your-secret-key

Linux/macOS:

export HUAWEICLOUD_SDK_AK="your-access-key"
export HUAWEICLOUD_SDK_SK="your-secret-key"

Note: Get your AK/SK from Huawei Cloud Console -> My Credentials -> Access Keys.

For complete environment variable configuration, see Environment Variables Guide.

Quick Start

1. Initialize a New Project

# Create a new agent project with LangGraph template
agentarts init -n my_agent -t langgraph

# Available templates: basic, langchain, langgraph, google-adk

This creates:

my_agent/
├── agent.py              # Agent implementation
├── requirements.txt      # Python dependencies
├── .agentarts_config.yaml # Project configuration
└── Dockerfile            # Docker build file

2. Configure Environment

Edit .agentarts_config.yaml to set environment variables:

runtime:
  environment_variables:
    - key: OPENAI_API_KEY
      value: "your-openai-api-key"
    - key: OPENAI_MODEL_NAME
      value: "gpt-4o-mini"  # Optional: gpt-4o, gpt-4-turbo, etc.
    - key: OPENAI_BASE_URL
      value: ""  # Optional: custom API endpoint

3. Local Development

# Start local development server
agentarts dev

# Server runs at http://127.0.0.1:8080
# Endpoints:
#   POST /invocations - Invoke agent
#   GET  /ping        - Health check

4. Deploy to Huawei Cloud

# Configure region
agentarts config set region cn-southwest-2

# Deploy to cloud
agentarts deploy

# Invoke deployed agent
agentarts invoke '{"message": "Hello, AgentArts!"}'

# Destroy deployment
agentarts destroy

CLI Commands Reference

Memory Commands

The agentarts memory command provides two groups of subcommands.

Space Management (requires AK/SK authentication):

Plugin Installation & Local Server:

Limitations & Requirements

Python Version

• Minimum: Python 3.10
• Recommended: Python 3.10 or 3.11

Framework Versions

When using optional framework dependencies, ensure the following minimum versions:

Note: LangGraph 1.0+ introduces a new Checkpoint format with required fields (step, pending_sends, parents). The SDK's integration module is compatible with LangGraph 1.0 and above.

Docker

Docker is required for:

• Building and deploying agents with agentarts deploy (alias: launch)

Install Docker:

• Docker Desktop for Windows
• Docker Engine for Linux

Resource Quotas

Refer to Huawei Cloud AgentArts Documentation for resource quotas and limits.

Documentation

• SDK User Guides - Memory, Code Interpreter, Gateway
• CLI User Guides - init, config, deploy, invoke, destroy
• Contributing Guide - Development setup and guidelines
• Architecture - System architecture overview

Development

# Install development dependencies
pip install -e ".[dev]"

# Run tests
pytest

# Code formatting
black . && isort .

# Type checking
mypy agentarts

# Linting
ruff check .

License

This project is licensed under the Apache 2.0 License - see the LICENSE file for details.

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for details.

Support

• Issues: GitHub Issues
• Documentation: https://support.huaweicloud.com/productdesc-agentarts/agentarts_03_0002.html
• Email: hwagentartssdk@huawei.com