AIOS: LLM Agent Operating System

The goal of AIOS is to build a Large Language Model (LLM) agent operating system, which intends to embed large language model into the operating system as the brain of the OS. AIOS is designed to address problems (e.g., scheduling, context switch, memory management, etc.) during the development and deployment of LLM-based agents, for a better ecosystem among agent developers and users.

🏠 Architecture of AIOS

AIOS provides the AIOS kernel as an abstraction on top of the OS kernel. The kernel facilitates the installation, execution and usage of agents. Furthermore, the AIOS SDK facilitates the development and deployment of agents.

📰 News

• [2024-09-01] 🔥 AIOS supports multiple agent creation frameworks (e.g., ReAct, Reflexion, OpenAGI, AutoGen, Open Interpreter, MetaGPT). Agents created by these frameworks can onboard AIOS. Onboarding guidelines can be found at the Doc.
• [2024-07-10] 📖 AIOS documentation is up, which can be found at Website.
• [2024-06-20] 🔥 Function calling for open-sourced LLMs (native huggingface, vllm, ollama) is supported.
• [2024-05-20] 🚀 More agents with ChatGPT-based tool calling are added (i.e., MathAgent, RecAgent, TravelAgent, AcademicAgent and CreationAgent), their profiles and workflows can be found in OpenAGI.
• [2024-05-13] 🛠️ Local models (diffusion models) as tools from HuggingFace are integrated.
• [2024-05-01] 🛠️ The agent creation in AIOS is refactored, which can be found in our OpenAGI package.
• [2024-04-05] 🛠️ AIOS currently supports external tool callings (google search, wolframalpha, rapid API, etc).
• [2024-04-02] 🤝 AIOS Discord Community is up. Welcome to join the community for discussions, brainstorming, development, or just random chats! For how to contribute to AIOS, please see CONTRIBUTE.
• [2024-03-25] ✈️ Our paper AIOS: LLM Agent Operating System is released!
• [2023-12-06] 📋 After several months of working, our perspective paper LLM as OS, Agents as Apps: Envisioning AIOS, Agents and the AIOS-Agent Ecosystem is officially released.

✈️ Getting Started

Please see our ongoing documentation for more information.

• Installation
• Quickstart
• WebUI Quickstart

Installation

Git clone AIOS

git clone https://github.com/agiresearch/AIOS.git
cd AIOS

Create venv environment (recommended)

python -m venv venv
source venv/bin/activate

or create conda environment

conda create -n venv python=3.10  # For Python 3.10
conda create -n venv python=3.11  # For Python 3.11
conda activate venv

If you have GPU environments, you can install the dependencies using

pip install -r requirements-cuda.txt

or else you can install the dependencies using

pip install -r requirements.txt

Quickstart

Tip

For the config of LLM endpoints, multiple API keys may be required to set up. Here we provide the .env.example to for easier configuration of these API keys, you can just copy .env.example as .env and set up the required keys based on your needs.

Note: Please use launch.py for the WebUI, or agent_repl.py for the TUI.

Use with OpenAI API

You need to get your OpenAI API key from https://platform.openai.com/api-keys. Then set up your OpenAI API key as an environment variable

export OPENAI_API_KEY=<YOUR_OPENAI_API_KEY>

Then run main.py with the models provided by OpenAI API

python main.py --llm_name gpt-3.5-turbo # use gpt-3.5-turbo for example

Use with Gemini API

You need to get your Gemini API key from https://ai.google.dev/gemini-api

export GEMINI_API_KEY=<YOUR_GEMINI_API_KEY>

Then run main.py with the models provided by OpenAI API

python main.py --llm_name gemini-1.5-flash # use gemini-1.5-flash for example

If you want to use open-sourced models provided by huggingface, here we provide three options:

• Use with ollama
• Use with native huggingface models
• Use with vllm

Use with ollama

You need to download ollama from from https://ollama.com/.

Then you need to start the ollama server either from ollama app

or using the following command in the terminal

ollama serve

To use models provided by ollama, you need to pull the available models from https://ollama.com/library

ollama pull llama3:8b # use llama3:8b for example

ollama can support CPU-only environment, so if you do not have CUDA environment

You can run aios with ollama models by

python main.py --llm_name ollama/llama3:8b --use_backend ollama # use ollama/llama3:8b for example

However, if you have the GPU environment, you can also pass GPU-related parameters to speed up using the following command

python main.py --llm_name ollama/llama3:8b --use_backend ollama --max_gpu_memory '{"0": "24GB"}' --eval_device "cuda:0" --max_new_tokens 256

Use with native huggingface llm models

Some of the huggingface models require authentification, if you want to use all of the models you need to set up your authentification token in https://huggingface.co/settings/tokens and set up it as an environment variable using the following command

export HF_AUTH_TOKENS=<YOUR_TOKEN_ID>

You can run with the

python main.py --llm_name meta-llama/Meta-Llama-3-8B-Instruct --max_gpu_memory '{"0": "24GB"}' --eval_device "cuda:0" --max_new_tokens 256

By default, huggingface will download the models in the ~/.cache directory. If you want to designate the download directory, you can set up it using the following command

export HF_HOME=<YOUR_HF_HOME>

Use with vllm

If you want to speed up the inference of huggingface models, you can use vllm as the backend.

Note

It is important to note that vllm currently only supports linux and GPU-enabled environment. So if you do not have the environment, you need to choose other options.

Considering that vllm itself does not support passing designated GPU ids, you need to either setup the environment variable,

export CUDA_VISIBLE_DEVICES="0" # replace with your designated gpu ids

Then run the command

python main.py --llm_name meta-llama/Meta-Llama-3-8B-Instruct --use_backend vllm --max_gpu_memory '{"0": "24GB"}' --eval_device "cuda:0" --max_new_tokens 256

or you can pass the CUDA_VISIBLE_DEVICES as the prefix

CUDA_VISIBLE_DEVICES=0 python main.py --llm_name meta-llama/Meta-Llama-3-8B-Instruct --use_backend vllm --max_gpu_memory '{"0": "24GB"}' --eval_device "cuda:0" --max_new_tokens 256

Web Quickstart

Requirements

Python

• Supported versions: Python 3.9 - 3.11

Node

• Supported versions: LTS support ONLY

you can check that you meet requirements by running

py -v

and

npm -v

in your terminal

Run the launch.py to start both the frontend and backend

python launch.py

which should open up https://localhost:3000 (if it doesn't, navigate to that on your browser)

Interact with all agents by using the @ to tag an agent.

Supported Agent Frameworks

• OpenAGI
• AutoGen
• Open-Interpreter
• MetaGPT

Supported LLM Endpoints

• OpenAI API
• Gemini API
• ollama
• vllm
• native huggingface models (locally)

🖋️ References

@article{mei2024aios,
  title={AIOS: LLM Agent Operating System},
  author={Mei, Kai and Li, Zelong and Xu, Shuyuan and Ye, Ruosong and Ge, Yingqiang and Zhang, Yongfeng}
  journal={arXiv:2403.16971},
  year={2024}
}
@article{ge2023llm,
  title={LLM as OS, Agents as Apps: Envisioning AIOS, Agents and the AIOS-Agent Ecosystem},
  author={Ge, Yingqiang and Ren, Yujie and Hua, Wenyue and Xu, Shuyuan and Tan, Juntao and Zhang, Yongfeng},
  journal={arXiv:2312.03815},
  year={2023}
}

🚀 Contributions

For how to contribute, see CONTRIBUTE. If you would like to contribute to the codebase, issues or pull requests are always welcome!

🌍 AIOS Contributors

🤝 Discord Channel

If you would like to join the community, ask questions, chat with fellows, learn about or propose new features, and participate in future developments, join our Discord Community!

📪 Contact

For issues related to AIOS development, we encourage submitting issues, pull requests, or initiating discussions in AIOS Discord Channel. For other issues please feel free to contact AIOS Foundation ([email protected]).