zabi-rs

The fastest, zero-allocation EVM ABI decoder for Rust.

zabi-rs is designed for high-performance applications (MEV bots, indexers, embedded devices) where every microsecond and byte of memory counts. It decodes EVM ABI data by validating and wrapping the raw specific byte slices, avoiding Vec and String allocations entirely.

⚡ Features

Zero Allocation: No heap allocations (malloc) during decoding.
no_std Compatible: Ready for embedded, WASM, and kernel-mode usage.
Safe & Fast: heavily audited unsafe pointer arithmetic wrapped in safe APIs.
Ecosystem Ready: Drop-in compatible with alloy-primitives types (optional).

🚀 Quick Start

[dependencies]
zabi-rs = "0.0.13"

use zabi_rs::{decode_tuple, ZU256, ZAddress};

fn main() -> Result<(), zabi_rs::ZError> {
    let data = [0u8; 64]; // Your ABI encoded data
    
    // Decode (uint256, address) tuple
    let (balance, owner) = decode_tuple!(&data, ZU256, ZAddress)?;
    
    println!("Balance: {:?}", balance);
    println!("Owner: {:?}", owner);
    
    Ok(())
}

🏎️ Performance

zabi-rs is orders of magnitude faster than traditional decoders because it skips the parsing/allocation phase.

Benchmark	zabi-rs	alloy-rs	ethers-rs
Uint256	~0.9 ns	~27 ns	~78 ns
Simple Tuple	~7.8 ns	~63 ns	~121 ns
Large Array	~2.5 ns	~831 ns	~5,170 ns

> Benchmarks run on M2 Air. See benches/ for details.

📚 Documentation

🗺️ Roadmap

Check out docs/roadmap.md for future plans (Zero-Copy Encoding, WASM).

License

MIT