Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/chaitu426/minibox/llms.txt

Use this file to discover all available pages before exploring further.

MiniBox stores images in an OCI-compatible blob layout under DataRoot (default /var/lib/minibox). These four commands let you inspect, remove, export, and import images managed by the daemon.

​
images

List all images registered in the daemon’s local index. 
minibox images [--json]
​
--json
boolean
default:"false"
Output the image list as a JSON array instead of the default formatted table.
# Table format
minibox images

# JSON format for scripting
minibox images --json
Each row shows the image name (as supplied to minibox build -t) and its total size in bytes — computed by summing the manifest blob, config blob, and all layer blobs from blobs/sha256/.
Only images that have the org.opencontainers.image.ref.name annotation set in index.json are listed. Intermediate or orphaned blobs not referenced by a named manifest are invisible to images and can be cleaned up with system prune.

​
rmi

Remove an image from the local index by its repository name. 
minibox rmi <image>
​
image
string
required
The image name exactly as registered at build time (e.g. myapp, webserver).
minibox rmi myapp
rmi removes the image’s entry from index.json. It does not immediately delete the underlying blobs from blobs/sha256/ — those are cleaned up lazily by system prune. This means rmi is safe to run even if another image shares blobs with the removed one.
After rmi, the image name can no longer be used with minibox run. Containers already running from this image are not affected — their overlay rootfs remains mounted until they exit.

​
save

Export a local image to a tar archive on the host filesystem. 
minibox save <image> <output.tar>
​
image
string
required
Name of the image to export (must exist in the local index).
​
output.tar
string
required
Destination path for the tar archive. The file is created or overwritten.
minibox save demo ./demo-image.tar
minibox save myapp /tmp/myapp-backup.tar

​
Archive format

The tar produced by save is not Docker-compatible. It is a minibox-specific archive containing:
FileDescription
meta.jsonJSON object with image (name) and manifest_digest fields
blobs/sha256/<digest>The image manifest blob
blobs/sha256/<digest>The image config blob
blobs/sha256/<digest>Each layer blob (gzip-compressed tar)
This format is designed for portability between minibox hosts, not for interoperability with Docker or other OCI runtimes.
Use save / load to transfer a built image to another machine without rebuilding. This is especially useful for images whose build context cannot be reproduced on the target host.

​
load

Import an image tar archive created by minibox save. 
minibox load <input.tar>
​
input.tar
string
required
Path to the tar archive previously created by minibox save.
minibox load ./demo-image.tar
load unpacks the tar archive under DataRoot, reads meta.json to determine the image name and manifest digest, and upserts the image reference into index.json. After a successful load, the image is immediately available for minibox run.

​
Round-trip example

# On the build host
minibox build -t demo .
minibox save demo /tmp/demo.tar
scp /tmp/demo.tar user@target-host:/tmp/

# On the target host
minibox load /tmp/demo.tar
minibox run -d demo
The load command merges blobs by content-addressed digest, so loading the same image twice is safe and idempotent.

Build docs developers (and LLMs) love

Get started for freeTalk to us