GAME Python SDK Library

The GAME Python SDK is a library that allows you interact and develop agents powered by the GAME architecture.

Overview of the Source Code Structure

There are 2 main folders in the source code (src folder)

1. game:

   Please refer to test_agent.py and test_worker.py for usage examples.

2. hosted_game
   

   This is a more abstracted version of the SDK which allows one to deploy a Twitter agent, which would then be hosted by GAME infrastructure.

   Note that this README focuses on the code in the game folder. More details about hosted_game can be found in src/hosted_game/README.md.

Installation

pip install game_sdk

To get the latest version of the SDK, clone the repository and install from source:

git clone https://github.com/game-by-virtuals/game-python.git
cd game-python
pip install -e .

To install the latest versions of the plugins, navigate to the plugin folder to run the installation, e.g.:

cd plugins/twitter
pip install -e .

Model Selection

The foundation models which power the GAME framework can also be confgiured and selected. This can be specified when creating agents or workers. The default model used is "Llama-3.1-405B-Instruct".

The models currently supported are:

• "Llama-3.1-405B-Instruct" (default)
• "Llama-3.3-70B-Instruct"
• "DeepSeek-R1"
• "DeepSeek-V3"
• "Qwen-2.5-72B-Instruct"

**Note: If model is not specified in the API call (REST API level) or on the SDK level when creating the agents, the default (Llama-3.1-405B-Instruct) model will be used.

Usage

1. game:

   • Request for a GAME API key in the Game Console https://console.game.virtuals.io/

   • Store the key in a safe location, like a .bashrc or a .zshrc file.

     export GAME_API_KEY="your_game_api_key"

   • Refer to src/game/README.md for usage guidelines, and refer to test_agent.py and test_worker.py for usage examples.

2. hosted_game
   

   • Open the Virtuals Platform and create/get an API key from the Agent Sandbox by clicking SDK/API Access

   • Store the key in a safe location, like a .bashrc or a .zshrc file.

     export VIRTUALS_API_KEY="your_virtuals_api_key"

   • Refer to src/hosted_game/README.md for usage examples.

If you have any trouble, contact Virtuals support or DevRel team members via Discord or Telegram

Examples

GAME

Example 1: test_agent.py

• Demonstrates how to create, configure, and run an AI agent with custom logic and workflows.

  python examples/game/test_agent.py

Example 2: test_worker.py

• Demonstrates how to configure, test, and manage workers that execute tasks within the agent's plan.

  python examples/game/test_worker.py

Example 3: example_twitter_reaction_module.py

• Demonstrates how to create and configure an AI agent with a reaction module for responding to tweets in real-time.

  python examples/game/example_twitter_reaction_module.py

Hosted Agent

Example 1: example-custom.py

• A complete guide to integrating an AI agent with other platform such as Telegram.

  python examples/hosted_agent/example-custom.py

Example 2: example-twitter.py

• A complete guide to integrating an AI agent with Twitter.

  python examples/hosted_agent/example-twitter.py

About G.A.M.E.

GAME is a modular agentic framework which enables an agent to plan actions and make decisions autonomously based on information provided to it.

Please refer to our whitepaper for more information and resources.

About GAME Python SDK

Currently, the SDK (specifically the code in the game folder) allows you to develop your agents powered by the GAME architecture in its most fullest and most flexible form.

The python SDK is made up of 3 main components (Agent, Worker, function), each with configurable arguments.

Agent (a.k.a. high level planner)

• Takes in a Goal
  • Drives the agents behaviour through the high level plan which influences the thinking and creation of tasks that would contribute towards this goal
• Takes in a Description
  • Combination of what was previously known as World Info + Agent Description
  • This include a description of the "world" the agent lives in, and the personality and background of the agent

Worker (a.k.a. low-level planner

• Takes in a Description
  • Used to control which workers are called by the agent, based on the high-level plan and tasks created to contribute to the goal

Function

• Takes in a Description
  • Used to control which functions are called by the workers, based on each worker's low-level plan
  • This can be any python executable

Features

• Develop your own custom agents for any application or platform.
• Ability to control your agents and workers via descriptions (prompts)
• Full control of what the agent sees (state) and can do (actions/functions)
• Ability to fully customise functions. This could include various combinations of programmed logic. For example:
  • Calling an API to retrieve data
  • Calling an API to retrieve data, followed by custom calculations or data processing logic in python code
  • 2 API calls chained together (e.g. calling an API to retrieve web data, and then posting a tweet)

ℹ️ Differences from the hosted GAME SDK version (i.e. hosted_game)

• Ability to fully customise functions (previously, each function was a single API call)
• Ability to control the low-level planner via description prompt (previously, only the high-level planner and functions could be controlled via description prompts)
• The description defined in the agent is equivalent to what was previously known as world information and agent description

How to Contribute

Want to help improve the project? Please see our detailed Contribution Guide.

Documentation

Detailed documentation to better understand the configurable components and the GAME architecture can be found on here.

Useful Resources

1. GAME TypeScript SDK

• This SDK mirrors the logic of the Python SDK but allows development in TypeScript.
• The repository contains TypeScript SDK documentation and contributed plugins .

2. Hosted GAME Agent

• This feature enables the configuration and deployment of a hosted agent that interacts with platforms like Twitter/X.
• Pre-built functions and actions are included for integration with the Twitter/X platform.
• It offers a similar experience to configuring an agent in the Agent Sandbox on the Virtuals Platform, but with a developer-friendly SDK interface.

3. GAME Whitepaper

• Comprehensive documentation about the GAME SDK and Sandbox.
• The How To section includes video tutorials, use cases, and additional tips.
• Content is regularly updated to provide new insights and improvements.