tigris bundle
Download multiple objects as a streaming tar archive in a single request. Designed for batch workloads that need many objects without per-object HTTP overhead.
Usage
tigris bundle <bucket> [flags]
t3 bundle <bucket> [flags]
Paths support the optional t3:// (or tigris://) prefix (e.g.
t3://my-bucket or just my-bucket).
Object keys can be provided via --keys (comma-separated list or path to a file
with one key per line) or piped through stdin.
Flags
| Name | Required | Default | Description |
|---|---|---|---|
--keys, -k | No | — | Comma-separated object keys, or path to a file with one key per line. If a local file matching the value exists, it is read as a keys file. If omitted, reads keys from stdin |
--output, -o | No | — | Output file path. Defaults to stdout (for piping) |
--compression | No | — | Compression algorithm for the archive. Auto-detected from output file extension when not specified. Options: none, gzip, zstd |
--on-error | No | skip | How to handle missing objects. Options: skip (omit them), fail (abort the request) |
Examples
# Bundle specific keys into a tar file
tigris bundle my-bucket --keys key1.jpg,key2.jpg --output archive.tar
# Read keys from a file
tigris bundle my-bucket --keys keys.txt --output archive.tar
# Compress with gzip
tigris bundle t3://my-bucket --keys keys.txt --compression gzip -o archive.tar.gz
# Pipe keys from stdin
cat keys.txt | tigris bundle my-bucket > archive.tar