vault-tar 0.1.1

AES-256-GCM file and directory encryption with chunked streaming and optional splitting.

vault-tar

AES-256-GCM file and directory encryption with chunked streaming, configurable compression, and optional output splitting.

Features

• AES-256-GCM authenticated encryption with PBKDF2-HMAC-SHA256 key derivation (1 200 000 iterations by default).
• Chunked streaming — encrypts/decrypts in 1 MiB chunks so memory usage stays constant regardless of input size.
• Compression — choose between xz, gz, bz2, zst (Python ≥ 3.14), or none.
• Output splitting — split encrypted output into fixed-size parts (default 1 GiB) for easier storage and transfer.
• Single-file mode — encrypt/decrypt individual files without tar archiving.
• Directory comparison — recursively compare two directory trees after a round-trip to verify integrity.
• Progress bars — optional verbose mode (-v) with real-time byte-level progress during compression, encryption, and decryption.

Requirements

• Python ≥ 3.12
• cryptography ≥ 43.0

Installation

From PyPI

pip install vault-tar

From source

git clone https://github.com/l1asis/vault-tar.git
cd vault-tar
pip install .

Usage

After installation the vtar command is available on your PATH.

Encrypt directories

vtar encrypt -i secret_docs/ photos/ -o encrypted -v

Encrypt with a specific algorithm

vtar encrypt -i data/ -o data --algorithm zst -v        # zstandard (Python ≥ 3.14)
vtar encrypt -i data/ -o data --algorithm gz -v          # gzip
vtar encrypt -i data/ -o data --algorithm none -v        # tar only, no compression

Encrypt a single file (skip archiving)

vtar encrypt -i backup.sql -o backup.enc --file-only --no-split

Decrypt

vtar decrypt -i encrypted -o restored/ -v
vtar decrypt -i backup.enc -o backup.sql --file-only

Compare directories

Verify that decrypted output matches the original:

vtar compare original/ restored/ -v

Additional options

Flag	Description
-v, --verbose	Show progress bars and status messages
--password TEXT	Supply password on command line (prompted if omitted)
--algorithm {xz,gz,bz2,zst,none}	Compression algorithm (default: xz)
--compression-level N	Algorithm-specific compression level
--chunk-size SIZE	Plaintext chunk size (default: 1MiB)
--split-size SIZE	Max part-file size (default: 1GiB)
--no-split	Write a single output file
--file-only	Encrypt/decrypt a single file directly
--cleanup	Remove original input after success
-V, --version	Show version and exit

Sizes accept human-readable suffixes: KiB, MiB, GiB.

File format

All encrypted output follows a custom binary format (v1):

Header (first part only):
  [2 B]  magic  0xEF01
  [1 B]  format version  0x01
  [2 B]  salt length          [N B]  salt
  [2 B]  base-nonce length    [12 B] base nonce

Chunks (sequential across parts):
  [4 B]  ciphertext length    [N B]  ciphertext (plaintext + 16 B GCM tag)

Each chunk uses a unique nonce derived as base_nonce XOR chunk_index (big-endian, 12 bytes) with AAD b"chunk_<index>".

Security

See SECURITY.md for the threat model, cryptographic details, known limitations, and responsible disclosure policy.

Acknowledgments

Built with assistance from GitHub Copilot (Claude Opus 4.6).

License

MIT — Copyright 2026 Volodymyr Horshenin