multibase

multibase is the Rust implementation of the multiformats multibase specification, a self-describing scheme that prefixes base-encoded data with a single character identifying which base was used. It supports the full multibase table including base16, base32, base58btc, base64, base45, and base256emoji.

Built for the IPFS/IPLD ecosystem, it offers a tiny, no_std-friendly API centered on encode and decode helpers, letting applications round-trip binary data through text encodings while preserving the base identifier so decoders never have to guess the format.

What You Get

• encode/decode functions covering the full multibase alphabet (base16, base32, base58btc, base64, base45, base256emoji, and more)
• A Base enum enumerating every supported encoding for type-safe selection
• no_std support via feature flags for embedded and constrained targets
• Self-describing output where the leading character records which base was used

Common Use Cases

• Encoding CIDs and hashes in IPFS/IPLD applications
• Round-tripping binary payloads through text-safe representations
• Decoding untrusted strings without knowing their base in advance

Under The Hood

Architecture - The crate is compact, with src/ split into base.rs (the Base enum and per-base metadata), encoding.rs, impls.rs, error.rs, and lib.rs exposing the public encode/decode functions. Each base variant maps to a concrete data-encoding alphabet, and the leading identifier character is read or written at the boundary of encode/decode.

Tech Stack - Rust (edition 2018) building on base-x, data-encoding, data-encoding-macro, base45, and base256emoji. std is a default feature that can be disabled for no_std; the README notes the data-encoding-macro dependency historically required nightly for the host_dep feature.

Code Quality - The repo includes a tests/ directory, criterion benches/ for performance, and a cli/ example, with typed errors in error.rs. The surface is small and well covered, and the README documents performance characteristics (base32/base64 are byte-aligned and fastest).

API Design - The API is deliberately minimal: multibase::encode(Base::Base64, bytes) and multibase::decode(str) returning (Base, Vec<u8>). The Base enum makes encoding selection type-safe and discoverable, giving a very shallow learning curve for anyone familiar with base encodings.