Installation

1. Clone the repository.

git clone git@github.com:oracle/agent-spec.git

Tip

If you face any problem, check with the Agent Spec team.

Next, install PyAgentSpec directly from source.

1. Create a fresh Python environment for building and running Agent Spec assistants:

  python3.10 -m venv <venv_name>
  source <venv_name>/bin/activate

2. Move to the agent-spec/pyagentspec directory:

  cd agent-spec/pyagentspec

3. Install pyagentspec:

  bash install-dev.sh

Extra dependencies

PyAgentSpec offers optional extra dependencies that can be installed to enable additional features.

• The autogen extra dependency gives access to the AutoGen runtime adapter.

• The langgraph extra dependency gives access to the LangGraph runtime adapter.

To install extra dependencies, run the following command specifying the list of dependencies you want to install:

pip install "pyagentspec[extra-dep-1,extra-dep-2]==26.1.0.dev0"

Supported platforms

PyAgentSpec strives for compatibility with major platforms and environments wherever possible.

Operating systems and CPU architectures

OS / CPU Architecture Support State	x86-64	ARM64
Linux	Supported	Untested
macOS	Supported	Supported

Python version

How to read the table:

• Unsupported: the package or one of its dependencies is not compatible with the Python version;

• Untested: the package and its dependencies are compatible with the Python version, but they are not tested;

• Supported: the package and its dependencies are compatible with the Python version, and the package is tested on that version.

Python version	Support State
Python 3.8	Unsupported
Python 3.9	Unsupported
Python 3.10	Supported
Python 3.11	Supported
Python 3.12	Supported
Python 3.13	Supported
Python 3.14	Supported

Package manager

Package Manager	Support State
pip	Supported
conda	Untested

Python implementation

Implementation	Support
CPython	Supported
PyPy	Untested

What do Supported, Untested and Unsupported mean?

• Unsupported: The package or one of its dependencies is not compatible with the Python version.

• Untested: The package and its dependencies are compatible with the Python version, but they are not tested.

• Supported: The package and its dependencies are compatible with the Python version, and the package is tested on that version.