payload-dumper-rust 🦀

A fast and efficient Android OTA payload dumper.

What it does

Extracts partition images (boot, system, vendor, etc.) from Android OTA payload.bin files.

This has these features

• Parallel extraction - Uses all CPU cores for faster extraction
• Works with ZIP files - Extract directly from ROM ZIPs without unzipping first
• URL support - Extract from remote URLs, downloading only the needed data instead of the entire file
• Cross-platform - Works on Linux, Windows, macOS, and Android (Termux)
• Incremental OTA - Experimental Incremental OTA Support ( Not tested)

Installation

Quick Install

Linux / macOS / Termux:

bash <(curl -sSL "https://raw.githubusercontent.com/rhythmcache/payload-dumper-rust/main/scripts/install.sh")

Windows (PowerShell):

powershell -NoExit -ExecutionPolicy Bypass -Command "Invoke-RestMethod -Uri 'https://raw.githubusercontent.com/rhythmcache/payload-dumper-rust/main/scripts/install.ps1' | Invoke-Expression"

Pre-built Binaries

Download from releases.

Build from Source

cargo install payload_dumper

Or:

git clone https://github.com/rhythmcache/payload-dumper-rust
cd payload-dumper-rust
cargo build --release

Usage

Basic Examples

Extract from payload.bin:

payload_dumper payload.bin -o output

Extract from ROM ZIP (no unzipping required):

payload_dumper rom.zip -o output

Extract from URL (downloads only required data):

payload_dumper https://example.com/ota.zip -o output

Extract specific partitions:

payload_dumper payload.bin -i boot,vendor_boot -o output

Extract Incremental OTA:

• To extract an incremental OTA, create a directory named old in the current working directory and copy all images from the previous build into it.
• If the old images are located elsewhere, specify their location using the --source-dir <PATH> argument.

List partitions without extracting:

payload_dumper payload.bin --list

Advanced Options

URL extraction with prefetch (better for slow connections):

payload_dumper https://example.com/ota.zip --prefetch -o output

Custom thread count:

payload_dumper payload.bin -t 8 -o output

Export metadata as JSON:

payload_dumper payload.bin --metadata -o output
# Creates output/payload_metadata.json

Full metadata with operation details:

payload_dumper payload.bin --metadata=full -o output

Skip verification (faster but not recommended):

payload_dumper payload.bin --no-verify -o output

All Options

Usage: payload_dumper [OPTIONS] <PAYLOAD_PATH>

Arguments:
  <PAYLOAD_PATH>  Path to payload.bin, ROM ZIP, or URL

Options:
  -o, --out <OUT>              Output directory [default: output]
  -U, --user-agent <AGENT>     Custom User-Agent for HTTP requests
  -C, --cookies <COOKIES>      Custom HTTP Cookie header value for remote requests
  -i, --images <IMAGES>        Comma-separated partition names to extract
  -t, --threads <THREADS>      Number of threads for parallel processing
  -l, --list                   List available partitions
  -m, --metadata[=<MODE>]      Save metadata as JSON (compact or full)
  -P, --no-parallel            Disable parallel extraction
  -n, --no-verify              Skip hash verification
      --prefetch               Download all data first (for remote URLs)
  -q, --quiet                  Suppress all non-essential output (errors will still be shown)
  -h, --help                   Show help
  -V, --version                Show version

Why extract from URLs?

Instead of downloading a 3GB OTA file to extract a 50MB boot image, this tool downloads only ~50-100MB of required data.

Useful for:

• Quick partition extraction without full downloads
• CI/CD pipelines
• Bandwidth-limited connections

Build

To build manually, you will need:

• Rust compiler and Cargo
• Basic command-line knowledge
• A bit of patience

1. Clone the repository

git clone --depth 1 https://github.com/rhythmcache/payload-dumper-rust.git
cd payload-dumper-rust

2. Default build (recommended for most users)

cargo build --release

This builds the binary with:

• Local .bin and .zip support
• Remote HTTP/HTTPS support
• Prefetch mode
• Metadata support

Feature-based builds

Payload Dumper is modular. You can disable features you do not need to:

• Reduce compile time
• Reduce binary size
• Avoid unnecessary dependencies

Local files only (fastest build, smallest binary)

If you only need local payload files and no network support:

cargo build --release --no-default-features

This:

• Disables all HTTP/network code
• Removes reqwest, DNS, and TLS dependencies
• Compiles much faster
• Produces a significantly smaller binary

Enable remote (HTTP/HTTPS) support

To work with remote URLs:

cargo build --release --features remote_zip

Static binaries and DNS notes

When building a fully static binary (for example with musl), DNS resolution may fail because system libc DNS resolvers are unavailable.

Use the built-in DNS resolver

cargo build --release --features hickory_dns

This enables:

• Built-in DNS resolution via hickory-resolver
• No dependency on system libc DNS
• Reliable behavior in fully static environments

By default, the resolver uses Cloudflare DNS (1.1.1.1).

Override DNS server at runtime

The DNS server is selected at runtime

To use a custom DNS server, set the environment variable when running the binary:

export PAYLOAD_DUMPER_CUSTOM_DNS=8.8.8.8,4.4.4.4
./payload_dumper

If you are unsure, use the default build.

Library usage

It also provides a few high level apis.

project using this library:

• payload-dumper-gui
  https://github.com/rhythmcache/payload-dumper-gui.git

Rust API

See

• src/extractor/local.rs

• src/extractor/remote.rs

C / C++ API convenience wrappers

• C API header:
  include/payload_dumper.h
• C++ wrapper header:
  include/payload_dumper.hpp

For API documentation and function behavior, refer directly to the headers.

You can regenerate C/C++ headers using cargo-c or cbindgen.

Troubleshooting

"Server doesn't support range requests"
Download the file first, then extract locally.

Out of memory errors
Use --no-parallel or reduce thread count with -t.

Slow extraction
Remove --no-parallel if set. For remote extraction, try --prefetch.

Credits

Inspired by vm03/payload_dumper

License

Apache-2.0