dwrs

A CLI file downloader with parallel downloading, resume support, and batch mode.

---

Installation

Via Cargo

cargo install dwrs

Build from source

git clone https://github.com/bircoder432/dwrs.git
cd dwrs
cargo build --release

Binary:

target/release/dwrs

---

Usage

Download a file

dwrs https://example.com/file.iso

Multiple files:

dwrs url1 url2 url3

---

Output files (IMPORTANT: repeated flag)

Output is specified per file using repeated -o flags:

dwrs url1 url2 -o out1 -o out2

Each -o corresponds to the URL at the same position.

If outputs are omitted, filenames are derived from URLs.

---

Download from file

Format:

https://example.com/file1.zip output1.zip
https://example.com/file2.zip
# comments supported

Run:

dwrs --file urls.txt

---

Resume download

dwrs https://example.com/file.zip --continue

---

Background mode

dwrs https://example.com/file.iso --background

---

Options

-w, --workers <N>              Parallel chunks per file (default: 4)
-o, --output <FILE>           Output filename (repeatable per URL)
-r, --retries <N>             Retry attempts (default: 3)
--buffer-size <KB>            Write buffer size in KB
--pool-size <N>               HTTP connection pool size (default: 100)
--max-files <N>               Max concurrent file downloads
--min-parallel-size <MB>     Minimum size for parallel download
-c, --continue               Resume interrupted downloads
-f, --file <PATH>            Input file with URL list
--config <PATH>              Custom config file
--background                 Run in background

---

Configuration

Default path:

~/.config/dwrs/config.toml

Example

workers = 8
buffer_size = 262144
pool_size = 100
retries = 3
min_parallel_size = 5242880

msg_template = "→ {output}"
template = "{spinner:.green} [{elapsed_precise}] {bar:40.cyan/blue} ({percent}%) {msg}"
bar_chars = "█▌░"

---

Behavior

Parallel download

• Uses HTTP Range requests
• Splits into chunks
• Downloads concurrently
• Merges final file
• Falls back to single-thread if unsupported

Resume

• Works for both sequential and parallel modes
• Requires server support for Range

---

Notes

• -o is positional by repetition, not a list
• URL → output mapping is index-based
• Parallel mode depends on server Accept-Ranges

---

License

Apache-2.0