Skip to content

armnet

Armnet is a managed arm farm for Physical AI researchers. Provision a robot and deploy a Docker container with a few lines of Python, and run on real hardware. armnet is the official Python SDK for interacting with the Armnet Arm Farm.

The Python SDK is split into three packages:

  • armnet-client: submit jobs, build/push images, manage secrets, and use volumes.
  • armnet-runtime: installed inside your container; provides @main, Context, and robot-cell helpers.
  • armnet-core: shared wire models and constants.

Quick Example

from armnet_client import Image, execute

image = Image.build(
    dockerfile="Dockerfile",
    context_dir=".",
    name="my-robot-runtime",
).push()

result = execute(
    image=image,
    embodiment="lerobot/so-101",
    task="assemble_block_tower",
    args={"calibrate": False},
    timeout_seconds=120,
)

print(result.return_value)

Local Docs

Preview these docs locally:

cd armnet
uv run --with mkdocs-material --with "mkdocstrings[python]" mkdocs serve

Build the static site:

cd armnet
uv run --with mkdocs-material --with "mkdocstrings[python]" mkdocs build --strict