How to Extract a Single File
Stream individual files from a remote image without downloading everything.
Prerequisites
- blobber installed
- Registry access configured
View File Contents
Stream a file to your terminal:
blobber cat ghcr.io/myorg/config:v1 app.yaml
Save to Local File
Use cp to copy directly to a file:
blobber cp ghcr.io/myorg/config:v1 app.yaml ./local-app.yaml
Or redirect cat output:
blobber cat ghcr.io/myorg/config:v1 app.yaml > local-app.yaml
Extract Nested Files
Use the full path as shown in blobber ls:
blobber cat ghcr.io/myorg/config:v1 config/database.yaml
Pipe to Other Commands
Parse JSON with jq
blobber cat ghcr.io/myorg/config:v1 settings.json | jq '.database'
Parse YAML with yq
blobber cat ghcr.io/myorg/config:v1 app.yaml | yq '.version'
Diff against local file
diff <(blobber cat ghcr.io/myorg/config:v1 app.yaml) ./local-app.yaml
Search with grep
blobber cat ghcr.io/myorg/config:v1 server.yaml | grep port
Extract Multiple Files
For multiple specific files, run multiple commands:
blobber cp ghcr.io/myorg/config:v1 app.yaml ./app.yaml
blobber cp ghcr.io/myorg/config:v1 database.yaml ./database.yaml
For many files, consider using blobber pull instead.
Binary Files
Binary files work the same way:
blobber cp ghcr.io/myorg/assets:v1 logo.png ./logo.png
Find the Correct Path
If you're unsure of the exact path:
blobber ls ghcr.io/myorg/config:v1 | grep database
Then use the exact path shown.
Why This Is Efficient
Blobber uses eStargz format with HTTP range requests. When you cat a file, only that file's bytes are downloaded, not the entire image.
For a 1GB image, extracting a 1KB config file downloads approximately 1KB.