Floresta

Floresta is a lightweight, embeddable Bitcoin client written in Rust. It does full block validation but replaces the UTXO set with Utreexo, a cryptographic accumulator that represents every unspent output as leaves in a small forest of Merkle trees. Combined with PoW Fraud Proofs and pruning, this keeps on-disk state under a gigabyte. In the Utreexo taxonomy (see BIP-0183), florestad runs as a Compact State Node: it validates every block and every script against the same consensus rules as Bitcoin Core but does not archive the full chain or keep the full UTXO set, so it can run on phones and single-board computers without giving up trust-minimization.

The project ships two main pieces: libfloresta, a set of reusable crates that applications can embed, and florestad, a daemon that runs as a standalone node with a watch-only wallet and a built-in Electrum server. Developers can use libfloresta the way they would use rust-bitcoin or bdk, building wallets, explorers, and indexers on top of a validating client. Rather than reimplementing a script interpreter, Floresta delegates script verification to libbitcoinkernel, the C++ library that exposes Bitcoin Core's validation engine. The project publishes an mdBook that walks through the codebase chapter by chapter.

Why fund it?

Verifying your own blocks and transactions is the strongest form of trust-minimization on Bitcoin. The cost of storing the full chain and UTXO set pushes most users onto custodial services or SPV-style light clients that trust third parties. Floresta does the same consensus work as a full archival node with a much smaller footprint, which makes it a practical fit for phones, single-board computers, and embedded wallet stacks.

OpenSats first funded Floresta in the seventh wave of Bitcoin grants in September 2024. The fourteenth wave added grants for contributors João Leal, Moises Pompilio, and Jose Castillo López. The project is open-source and available at getfloresta/Floresta, with origins at Vinteum, the Brazilian nonprofit that incubated it.

What's next?

Floresta v0.9.0 shipped in April 2026 (announcement post). On the wire, the node speaks BIP-0183 Utreexo messaging for proof-driven sync. Script validation moved from the deprecated libbitcoinconsensus to libbitcoinkernel, which runs roughly 15x faster. The JSON-RPC picked up rescanblockchain, gettxout, ping, addnode, and JSON-RPC v1 for legacy clients, so most Bitcoin Core tooling works against florestad unchanged. v0.9.0 also landed a modular mempool (the groundwork for transaction relay), parallel pytest functional tests, more fuzz targets, no_std portability work, and dual licensing under MIT OR Apache-2.0. For reference, v0.8.0 (July 2025) added BIP-0324 P2P v2, testnet4, Prometheus metrics, and the first mdBook, while v0.8.1 (January 2026) was a bugfix release covering reindex and filter issues.

The next big item is SwiftSync. It pairs implicit deletion in the Utreexo accumulator with pre-generated hints so blocks can be validated out of order without downloading proofs. In one test, syncing signet to block 290,000 on a home connection went from 1h 15min with plain Utreexo to 6 minutes with witnessless SwiftSync. The Utreexo BIP draft is under review. Other open work includes evaluating BDK as the watch-only wallet, packaging florestad for nix-bitcoin and upstream nixpkgs, transaction relay on top of the new mempool, and more P2P integration tests.