Overview

Introduction

fhEVM is a protocol enabling confidential smart contracts on EVM-compatible blockchains. By leveraging Fully Homomorphic Encryption (FHE), fhEVM ensures complete data privacy without sacrificing composability or usability.

Core principles

The design of fhEVM is guided by the following principles:

1. Preserving security: No impact on the underlying blockchain's security guarantees.

2. Public verifiability: All computations are publicly verifiable while keeping data confidential.

3. Developer accessibility: Build confidential smart contracts using familiar Solidity tooling, without requiring cryptographic expertise.

4. Composability: Confidential smart contracts are fully interoperable with each other and public contracts.

Key features

Encrypted data types

fhEVM introduces encrypted data types compatible with Solidity:

• Booleans: ebool

• Unsigned Integers: euint4, euint8, euint16, euint32, euint64, euint128, euint256

• Addresses: eaddress

• Bytes: ebytes64, ebytes128, ebytes256

• Input: einput for handling encrypted input data

Encrypted data is represented as ciphertext handles, ensuring secure computation and interaction.

For more information see use of encrypted types.

Casting types

fhEVM provides functions to cast between encrypted types:

• Casting between encrypted types: TFHE.asEbool converts encrypted integers to encrypted booleans

• Casting to encrypted types: TFHE.asEuintX converts plaintext values to encrypted types

• Casting to encrypted addresses: TFHE.asEaddress converts plaintext addresses to encrypted addresses

• Casting to encrypted bytes: TFHE.asEbytesX converts plaintext bytes to encrypted bytes

For more information see use of encrypted types.

Confidential computation

fhEVM enables symbolic execution of encrypted operations, supporting:

• Arithmetic: TFHE.add, TFHE.sub, TFHE.mul, TFHE.min, TFHE.max, TFHE.neg, TFHE.div, TFHE.rem

  • Note: div and rem operations are supported only with plaintext divisors

• Bitwise: TFHE.and, TFHE.or, TFHE.xor, TFHE.not, TFHE.shl, TFHE.shr, TFHE.rotl, TFHE.rotr

• Comparison: TFHE.eq, TFHE.ne, TFHE.lt, TFHE.le, TFHE.gt, TFHE.ge

• Advanced: TFHE.select for branching on encrypted conditions, TFHE.randEuintX for on-chain randomness.

For more information on operations, see Operations on encrypted types.

For more information on conditional branching, see Conditional logic in FHE.

For more information on random number generation, see Generate Random Encrypted Numbers.

Access control mechanism

fhEVM enforces access control with a blockchain-based Access Control List (ACL):

• Persistent access: TFHE.allow, TFHE.allowThis grants permanent permissions for ciphertexts.

• Transient access: TFHE.allowTransient provides temporary access for specific transactions.

• Validation: TFHE.isSenderAllowed ensures that only authorized entities can interact with ciphertexts.

For more information see ACL.

Architectural overview

The fhEVM architecture provides the foundation for confidential smart contracts on EVM-compatible blockchains. At its core is FHE, a cryptographic technique enabling computations directly on encrypted data, ensuring privacy at every stage.

This system relies on three key types:

• The public key: used for encrypting data.

• The private key: used for decryption and securely managed by the Key Management System or KMS

• The evaluation key: enabling encrypted computations performed by the coprocessor.

The fhEVM leverages Zama's TFHE library, integrating seamlessly with blockchain environments to address transparency, composability, and scalability challenges. Its hybrid architecture combines:

• On-chain smart contracts for encrypted state management and access controls.

• Off-chain coprocessors for resource-intensive FHE computations.

• The Gateway to coordinate between blockchain, KMS, and coprocessors.

• The KMS for secure cryptographic key management and proof validation.

This architecture enables developers to write private, composable smart contracts using symbolic execution and zero-knowledge proofs, ensuring data confidentiality and computational integrity.

For a detailed exploration of the fhEVM architecture, including components, workflows, and deployment models, see Architecture Overview.