Run captured programs in a Docker container you started.
The docker machine provider reaches a running container, by name, and
runs the programs it captures in it. You start
and stop the container; Mirage only runs programs in it. Python SDK and
Node.
Installation
Python
TypeScript
pip install 'mirage-ai[ssh]'
pnpm add ssh2
These serve the workspace into the container with mount_type: ssh. Start
the container with FUSE allowed and sshfs in it, or from an image that has
what connecting the workspace needs:
python3 runs in the container and reads the file Mirage wrote, so this
prints HELLO. Remove both with mirage workspace delete box-demo and
docker rm -f my-sandbox.
Connecting the workspace
Pick a mount_type. With ssh, the image needs only sshfs. With
mirage, it needs fuse3 and uv or npm to install
mirage-fuse. With own,
copy sandbox.yaml in and start Mirage in the container:
workspace.pause() closes the workspace, then runs docker pause: the
container’s processes are frozen and use no CPU, but it keeps its memory. The
next line’s connect runs docker unpause. Without on_pause,
workspace.pause() leaves the container running unless the call names the
machine. See Pause.
Notes
container takes a name or an ID. The Docker CLI is the transport, so
Docker Desktop, Colima and a compatible Podman alias all work.
Sizing, GPUs, networks, volumes and bind mounts go on your own
docker run. You can bake sandbox.yaml into the image instead of
copying it, and point it at host.docker.internal for a host service.
Environment values go on the docker exec command line, so other users of
the host can see them in ps. Avoid secrets on a host shared with
untrusted users.