Overview
The library exposes utility functions for FHE operations. The goal of the library is to provide a seamless developer experience for writing smart contracts that can operate on confidential data.
Types
The library provides a type system that is checked both at compile time and at run time. The structure and operations related to these types are described in this section.
We currently support encrypted integers of bit length up to 128 bits and special types such as ebool and eaddress.
The encrypted integers behave as much as possible as Solidity’s integer types. However, behavior such as “revert on overflow” is not supported as this would leak some information about the encrypted integers. Therefore, arithmetic on euint types is unchecked, i.e. there is wrap-around on overflow.
In the back-end, encrypted integers are FHE ciphertexts. The library abstracts away the ciphertexts and presents pointers to ciphertexts, or ciphertext handles, to the smart contract developer. The euint, ebool and eaddress types are wrappers over these handles.
Supported Types
Compute Types
| Name | Bit Size | Usage |
|---|
euint8 | 8 | Compute |
euint16 | 16 | Compute |
euint32 | 32 | Compute |
euint64 | 64 | Compute |
euint128 | 128 | Compute |
ebool | 8 | Compute |
eaddress | 160 | Compute |
Input Types
| Name | Bit Size | Usage |
|---|
InEuint8 | 8 | Input |
InEuint16 | 16 | Input |
InEuint32 | 32 | Input |
InEuint64 | 64 | Input |
InEuint128 | 128 | Input |
InEbool | 8 | Input |
InEaddress | 160 | Input |
The ebool type is not a real boolean type. It is implemented as a euint8 for compatibility with FHE operations.
Operations
There are two ways to perform operations with FHE.sol:
Using Direct Function Calls
Direct function calls are the most straightforward way to perform operations with FHE.sol. For example, if you want to add two encrypted 8-bit integers (euint8), you can do so as follows:
euint8 result = FHE.add(lhs, rhs);
lhs and rhs are your euint8 variables, and result will store the outcome of the addition.
Using Library Bindings
FHE.sol also provides library bindings, allowing for a more natural syntax. To use this, you first need to include the library for your specific data type. For euint8, the usage would look like this:
euint8 result = lhs.add(rhs);
lhs.add(rhs) performs the addition using the library function implicitly.
Supported Operations
Complete documentation of every function in FHE.sol (including inputs and outputs) can be found in the FHE.sol API Reference. | Name | FHE.sol function | Operator | euint8 | euint16 | euint32 | euint64 | euint128 | ebool | eaddress |
|---|
| Addition | add | + | Yes | Yes | Yes | Yes | Yes | No | No |
| Subtraction | sub | - | Yes | Yes | Yes | Yes | Yes | No | No |
| Multiplication | mul | * | Yes | Yes | Yes | Yes | Yes | No | No |
| Bitwise And | and | & | Yes | Yes | Yes | Yes | Yes | Yes | No |
| Bitwise Or | or | | | Yes | Yes | Yes | Yes | Yes | Yes | No |
| Bitwise Xor | xor | ^ | Yes | Yes | Yes | Yes | Yes | Yes | No |
| Division | div | / | Yes | Yes | Yes | Yes | Yes | No | No |
| Remainder | rem | % | Yes | Yes | Yes | Yes | Yes | No | No |
| Square | square | n/a | Yes | Yes | Yes | Yes | Yes | No | No |
| Shift Right | shr | n/a | Yes | Yes | Yes | Yes | Yes | No | No |
| Shift Left | shl | n/a | Yes | Yes | Yes | Yes | Yes | No | No |
| Rotate Right | ror | n/a | Yes | Yes | Yes | Yes | Yes | No | No |
| Rotate Left | rol | n/a | Yes | Yes | Yes | Yes | Yes | No | No |
| Equal | eq | n/a | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Not equal | ne | n/a | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Greater than or equal | gte | n/a | Yes | Yes | Yes | Yes | Yes | No | No |
| Greater than | gt | n/a | Yes | Yes | Yes | Yes | Yes | No | No |
| Less than or equal | lte | n/a | Yes | Yes | Yes | Yes | Yes | No | No |
| Less than | lt | n/a | Yes | Yes | Yes | Yes | Yes | No | No |
| Min | min | n/a | Yes | Yes | Yes | Yes | Yes | No | No |
| Max | max | n/a | Yes | Yes | Yes | Yes | Yes | No | No |
| Not | not | n/a | Yes | Yes | Yes | Yes | Yes | Yes | No |
| Select | select | n/a | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Decrypt | decrypt | n/a | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
Division and Remainder by 0: These operations will output an encrypted representation of the maximal value of the uint type being used (e.g., encrypted 255 for euint8). This behavior prevents information leakage about the divisor.
Operation Examples
Arithmetic Operations
// Addition
euint32 sum = FHE.add(a, b);
// or using library binding
euint32 sum = a.add(b);
// Multiplication
euint32 product = FHE.mul(a, b);
Comparison Operations
// Returns an encrypted boolean (ebool)
ebool isGreater = FHE.gt(amount, threshold);
ebool isEqual = FHE.eq(value1, value2);
Bitwise Operations
// Bitwise AND
euint16 masked = FHE.and(flags, mask);
// Bitwise shifts
euint32 shifted = FHE.shl(value, 4); // Shift left by 4 bits
Control Flow
// Select between two values based on encrypted condition
euint32 result = FHE.select(condition, valueIfTrue, valueIfFalse);
