Resilience Network
Pre-mainnet research network No public testnet yet Protocol manifest r28 Shielded transfers experimental Current status

Protocol research, pre-mainnet

An experimental proof‑of‑stake network, documented as it is.

Resilience Network is a proof-of-stake blockchain written in Go and Rust. It combines BFT finality and secret leader election with an EVM execution layer, native account authorization, and Orchard shielded transfers. This site publishes its specifications, its evidence and its open problems, each tied to a specific revision.

Read the protocol See current status

Configured defaults

Values from the reference configuration. They are targets and limits, not measured guarantees; measured behavior is on Assurance.

2 s
target slot duration
192
slots per epoch, 384 s at target
128
maximum active validators
>2/3
voting power to finalize
67
registered JSON-RPC methods
70
live-network user stories

What the protocol does

Six feature families, each with a specification page. The label says whether it is in the reference implementation, behind an activation switch, or still experimental.

  • BFT finality, secret leaders

    Durable locks and finality certificates. Proposers are chosen privately with an RFC 9381 verifiable random function, so nobody can read the schedule in advance.

    Implemented

  • EVM execution

    An EVM with Shanghai semantics, differentially tested against go-ethereum. Design deviations are written down and pinned by tests.

    Implemented

  • Native account authorization

    Accounts keep one address while authority rotates: threshold root keys, expiring policy-bound session keys, P-256 passkey sessions.

    Implemented

  • Recovery, sponsorship, names

    Delayed guardian recovery, sponsored gas, and .rsc names registered by commit and reveal.

    ImplementedNames need activation

  • Delegation shares

    Validator-specific, non-rebasing delegation shares that can change hands, with a read-only vault interface for contracts.

    ImplementedNeeds activation

  • Orchard shielded transfers

    Real Halo 2 proofs verified in consensus, with no trusted setup. Off for the first public testnet until an external audit.

    Experimental

What this is not, yet

Five things people reasonably assume about a new chain that are not true of this one today.

Not live

There is no public testnet or mainnet. Development networks run on test hardware.

Not externally audited

Every review to date is internal. The shielded layer must pass an independent audit before it handles value.

Not Ethereum wire-compatible

RPC method names are familiar, but transactions use the chain's own V2 format. Ethereum wallets cannot sign them.

Not a token offering

RSC is the protocol's native unit for fees, staking and rewards. Nothing here is an offer or a price.

Not a finality promise

Two seconds is a slot target. Measured behavior under load is published on Assurance.

Where to start

Suggested reading order by what you came to do.

Protocol researchersConsensus, then Privacy, then Decisions Prospective validatorsStatus, then Network and economics, then load results Wallet and dApp developersDevelopers, then Accounts, then Execution Security reviewersAssurance, then Security and disclosure

Latest from the log

Protocol revisions, reviews and measurements, newest first. Read the full log.

DateEntry
2026-10-13Internal security review of ef16072, dated 2026-10-12, added to the record. Open availability findings block the public testnet.
2026-10-09Public testnet launch procedure defined: transparent-only, shielded transfers off.
2026-10-08Manifest r28. Delayed consensus rounds now catch up from authenticated peer progress.